]> git.proxmox.com Git - mirror_ubuntu-focal-kernel.git/blame - drivers/net/myri10ge/myri10ge.c
Fix common misspellings
[mirror_ubuntu-focal-kernel.git] / drivers / net / myri10ge / myri10ge.c
CommitLineData
0da34b6d
BG
1/*************************************************************************
2 * myri10ge.c: Myricom Myri-10G Ethernet driver.
3 *
e3fd5534 4 * Copyright (C) 2005 - 2009 Myricom, Inc.
0da34b6d
BG
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of Myricom, Inc. nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
18 *
4a2e612a
BG
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
20 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
0da34b6d 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
4a2e612a
BG
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
23 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29 * POSSIBILITY OF SUCH DAMAGE.
0da34b6d
BG
30 *
31 *
32 * If the eeprom on your board is not recent enough, you will need to get a
33 * newer firmware image at:
34 * http://www.myri.com/scs/download-Myri10GE.html
35 *
36 * Contact Information:
37 * <help@myri.com>
38 * Myricom, Inc., 325N Santa Anita Avenue, Arcadia, CA 91006
39 *************************************************************************/
40
78ca90ea
JP
41#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
42
0da34b6d
BG
43#include <linux/tcp.h>
44#include <linux/netdevice.h>
45#include <linux/skbuff.h>
46#include <linux/string.h>
47#include <linux/module.h>
48#include <linux/pci.h>
b10c0668 49#include <linux/dma-mapping.h>
0da34b6d
BG
50#include <linux/etherdevice.h>
51#include <linux/if_ether.h>
52#include <linux/if_vlan.h>
1e6e9342 53#include <linux/inet_lro.h>
981813d8 54#include <linux/dca.h>
0da34b6d
BG
55#include <linux/ip.h>
56#include <linux/inet.h>
57#include <linux/in.h>
58#include <linux/ethtool.h>
59#include <linux/firmware.h>
60#include <linux/delay.h>
0da34b6d
BG
61#include <linux/timer.h>
62#include <linux/vmalloc.h>
63#include <linux/crc32.h>
64#include <linux/moduleparam.h>
65#include <linux/io.h>
199126a2 66#include <linux/log2.h>
5a0e3ad6 67#include <linux/slab.h>
0da34b6d 68#include <net/checksum.h>
1e6e9342
AG
69#include <net/ip.h>
70#include <net/tcp.h>
0da34b6d
BG
71#include <asm/byteorder.h>
72#include <asm/io.h>
0da34b6d
BG
73#include <asm/processor.h>
74#ifdef CONFIG_MTRR
75#include <asm/mtrr.h>
76#endif
77
78#include "myri10ge_mcp.h"
79#include "myri10ge_mcp_gen_header.h"
80
2a3f2790 81#define MYRI10GE_VERSION_STR "1.5.2-1.459"
0da34b6d
BG
82
83MODULE_DESCRIPTION("Myricom 10G driver (10GbE)");
84MODULE_AUTHOR("Maintainer: help@myri.com");
85MODULE_VERSION(MYRI10GE_VERSION_STR);
86MODULE_LICENSE("Dual BSD/GPL");
87
88#define MYRI10GE_MAX_ETHER_MTU 9014
89
90#define MYRI10GE_ETH_STOPPED 0
91#define MYRI10GE_ETH_STOPPING 1
92#define MYRI10GE_ETH_STARTING 2
93#define MYRI10GE_ETH_RUNNING 3
94#define MYRI10GE_ETH_OPEN_FAILED 4
95
96#define MYRI10GE_EEPROM_STRINGS_SIZE 256
97#define MYRI10GE_MAX_SEND_DESC_TSO ((65536 / 2048) * 2)
1e6e9342
AG
98#define MYRI10GE_MAX_LRO_DESCRIPTORS 8
99#define MYRI10GE_LRO_MAX_PKTS 64
0da34b6d 100
40f6cff5 101#define MYRI10GE_NO_CONFIRM_DATA htonl(0xffffffff)
0da34b6d
BG
102#define MYRI10GE_NO_RESPONSE_RESULT 0xffffffff
103
dd50f336
BG
104#define MYRI10GE_ALLOC_ORDER 0
105#define MYRI10GE_ALLOC_SIZE ((1 << MYRI10GE_ALLOC_ORDER) * PAGE_SIZE)
106#define MYRI10GE_MAX_FRAGS_PER_FRAME (MYRI10GE_MAX_ETHER_MTU/MYRI10GE_ALLOC_SIZE + 1)
107
236bb5e6
BG
108#define MYRI10GE_MAX_SLICES 32
109
0da34b6d 110struct myri10ge_rx_buffer_state {
dd50f336
BG
111 struct page *page;
112 int page_offset;
c755b4b6
FT
113 DEFINE_DMA_UNMAP_ADDR(bus);
114 DEFINE_DMA_UNMAP_LEN(len);
0da34b6d
BG
115};
116
117struct myri10ge_tx_buffer_state {
118 struct sk_buff *skb;
119 int last;
c755b4b6
FT
120 DEFINE_DMA_UNMAP_ADDR(bus);
121 DEFINE_DMA_UNMAP_LEN(len);
0da34b6d
BG
122};
123
124struct myri10ge_cmd {
125 u32 data0;
126 u32 data1;
127 u32 data2;
128};
129
130struct myri10ge_rx_buf {
131 struct mcp_kreq_ether_recv __iomem *lanai; /* lanai ptr for recv ring */
0da34b6d
BG
132 struct mcp_kreq_ether_recv *shadow; /* host shadow of recv ring */
133 struct myri10ge_rx_buffer_state *info;
dd50f336
BG
134 struct page *page;
135 dma_addr_t bus;
136 int page_offset;
0da34b6d 137 int cnt;
dd50f336 138 int fill_cnt;
0da34b6d
BG
139 int alloc_fail;
140 int mask; /* number of rx slots -1 */
dd50f336 141 int watchdog_needed;
0da34b6d
BG
142};
143
144struct myri10ge_tx_buf {
145 struct mcp_kreq_ether_send __iomem *lanai; /* lanai ptr for sendq */
236bb5e6
BG
146 __be32 __iomem *send_go; /* "go" doorbell ptr */
147 __be32 __iomem *send_stop; /* "stop" doorbell ptr */
0da34b6d
BG
148 struct mcp_kreq_ether_send *req_list; /* host shadow of sendq */
149 char *req_bytes;
150 struct myri10ge_tx_buffer_state *info;
151 int mask; /* number of transmit slots -1 */
0da34b6d
BG
152 int req ____cacheline_aligned; /* transmit slots submitted */
153 int pkt_start; /* packets started */
b53bef84
BG
154 int stop_queue;
155 int linearized;
0da34b6d
BG
156 int done ____cacheline_aligned; /* transmit slots completed */
157 int pkt_done; /* packets completed */
b53bef84 158 int wake_queue;
236bb5e6 159 int queue_active;
0da34b6d
BG
160};
161
162struct myri10ge_rx_done {
163 struct mcp_slot *entry;
164 dma_addr_t bus;
165 int cnt;
166 int idx;
1e6e9342
AG
167 struct net_lro_mgr lro_mgr;
168 struct net_lro_desc lro_desc[MYRI10GE_MAX_LRO_DESCRIPTORS];
0da34b6d
BG
169};
170
b53bef84
BG
171struct myri10ge_slice_netstats {
172 unsigned long rx_packets;
173 unsigned long tx_packets;
174 unsigned long rx_bytes;
175 unsigned long tx_bytes;
176 unsigned long rx_dropped;
177 unsigned long tx_dropped;
178};
179
180struct myri10ge_slice_state {
0da34b6d
BG
181 struct myri10ge_tx_buf tx; /* transmit ring */
182 struct myri10ge_rx_buf rx_small;
183 struct myri10ge_rx_buf rx_big;
184 struct myri10ge_rx_done rx_done;
b53bef84
BG
185 struct net_device *dev;
186 struct napi_struct napi;
187 struct myri10ge_priv *mgp;
188 struct myri10ge_slice_netstats stats;
189 __be32 __iomem *irq_claim;
190 struct mcp_irq_data *fw_stats;
191 dma_addr_t fw_stats_bus;
192 int watchdog_tx_done;
193 int watchdog_tx_req;
d0234215 194 int watchdog_rx_done;
5dd2d332 195#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
196 int cached_dca_tag;
197 int cpu;
198 __be32 __iomem *dca_tag;
199#endif
0dcffac1 200 char irq_desc[32];
b53bef84
BG
201};
202
203struct myri10ge_priv {
0dcffac1 204 struct myri10ge_slice_state *ss;
b53bef84 205 int tx_boundary; /* boundary transmits cannot cross */
0dcffac1 206 int num_slices;
b53bef84
BG
207 int running; /* running? */
208 int csum_flag; /* rx_csums? */
0da34b6d 209 int small_bytes;
dd50f336 210 int big_bytes;
fa0a90d9 211 int max_intr_slots;
0da34b6d 212 struct net_device *dev;
b53bef84 213 spinlock_t stats_lock;
0da34b6d
BG
214 u8 __iomem *sram;
215 int sram_size;
216 unsigned long board_span;
217 unsigned long iomem_base;
40f6cff5 218 __be32 __iomem *irq_deassert;
0da34b6d
BG
219 char *mac_addr_string;
220 struct mcp_cmd_response *cmd;
221 dma_addr_t cmd_bus;
0da34b6d
BG
222 struct pci_dev *pdev;
223 int msi_enabled;
0dcffac1
BG
224 int msix_enabled;
225 struct msix_entry *msix_vectors;
5dd2d332 226#ifdef CONFIG_MYRI10GE_DCA
981813d8 227 int dca_enabled;
ef09aadf 228 int relaxed_order;
981813d8 229#endif
66341fff 230 u32 link_state;
0da34b6d
BG
231 unsigned int rdma_tags_available;
232 int intr_coal_delay;
40f6cff5 233 __be32 __iomem *intr_coal_delay_ptr;
0da34b6d 234 int mtrr;
276e26c3 235 int wc_enabled;
0da34b6d
BG
236 int down_cnt;
237 wait_queue_head_t down_wq;
238 struct work_struct watchdog_work;
239 struct timer_list watchdog_timer;
0da34b6d 240 int watchdog_resets;
b53bef84 241 int watchdog_pause;
0da34b6d 242 int pause;
7d351035 243 bool fw_name_allocated;
0da34b6d
BG
244 char *fw_name;
245 char eeprom_strings[MYRI10GE_EEPROM_STRINGS_SIZE];
c0bf8801 246 char *product_code_string;
0da34b6d 247 char fw_version[128];
9dc6f0e7
BG
248 int fw_ver_major;
249 int fw_ver_minor;
250 int fw_ver_tiny;
251 int adopted_rx_filter_bug;
0da34b6d
BG
252 u8 mac_addr[6]; /* eeprom mac address */
253 unsigned long serial_number;
254 int vendor_specific_offset;
85a7ea1b 255 int fw_multicast_support;
04ed3e74 256 u32 features;
4f93fde0 257 u32 max_tso6;
0da34b6d
BG
258 u32 read_dma;
259 u32 write_dma;
260 u32 read_write_dma;
c58ac5ca
BG
261 u32 link_changes;
262 u32 msg_enable;
2d90b0aa 263 unsigned int board_number;
d0234215 264 int rebooted;
0da34b6d
BG
265};
266
267static char *myri10ge_fw_unaligned = "myri10ge_ethp_z8e.dat";
268static char *myri10ge_fw_aligned = "myri10ge_eth_z8e.dat";
0dcffac1
BG
269static char *myri10ge_fw_rss_unaligned = "myri10ge_rss_ethp_z8e.dat";
270static char *myri10ge_fw_rss_aligned = "myri10ge_rss_eth_z8e.dat";
b9721d5a
BH
271MODULE_FIRMWARE("myri10ge_ethp_z8e.dat");
272MODULE_FIRMWARE("myri10ge_eth_z8e.dat");
273MODULE_FIRMWARE("myri10ge_rss_ethp_z8e.dat");
274MODULE_FIRMWARE("myri10ge_rss_eth_z8e.dat");
0da34b6d 275
7d351035 276/* Careful: must be accessed under kparam_block_sysfs_write */
0da34b6d
BG
277static char *myri10ge_fw_name = NULL;
278module_param(myri10ge_fw_name, charp, S_IRUGO | S_IWUSR);
d1ce3a0f 279MODULE_PARM_DESC(myri10ge_fw_name, "Firmware image name");
0da34b6d 280
2d90b0aa
BG
281#define MYRI10GE_MAX_BOARDS 8
282static char *myri10ge_fw_names[MYRI10GE_MAX_BOARDS] =
7fe624f5 283 {[0 ... (MYRI10GE_MAX_BOARDS - 1)] = NULL };
2d90b0aa
BG
284module_param_array_named(myri10ge_fw_names, myri10ge_fw_names, charp, NULL,
285 0444);
286MODULE_PARM_DESC(myri10ge_fw_name, "Firmware image names per board");
287
0da34b6d
BG
288static int myri10ge_ecrc_enable = 1;
289module_param(myri10ge_ecrc_enable, int, S_IRUGO);
d1ce3a0f 290MODULE_PARM_DESC(myri10ge_ecrc_enable, "Enable Extended CRC on PCI-E");
0da34b6d 291
0da34b6d
BG
292static int myri10ge_small_bytes = -1; /* -1 == auto */
293module_param(myri10ge_small_bytes, int, S_IRUGO | S_IWUSR);
d1ce3a0f 294MODULE_PARM_DESC(myri10ge_small_bytes, "Threshold of small packets");
0da34b6d
BG
295
296static int myri10ge_msi = 1; /* enable msi by default */
3621cec5 297module_param(myri10ge_msi, int, S_IRUGO | S_IWUSR);
d1ce3a0f 298MODULE_PARM_DESC(myri10ge_msi, "Enable Message Signalled Interrupts");
0da34b6d 299
f761fae1 300static int myri10ge_intr_coal_delay = 75;
0da34b6d 301module_param(myri10ge_intr_coal_delay, int, S_IRUGO);
d1ce3a0f 302MODULE_PARM_DESC(myri10ge_intr_coal_delay, "Interrupt coalescing delay");
0da34b6d
BG
303
304static int myri10ge_flow_control = 1;
305module_param(myri10ge_flow_control, int, S_IRUGO);
d1ce3a0f 306MODULE_PARM_DESC(myri10ge_flow_control, "Pause parameter");
0da34b6d
BG
307
308static int myri10ge_deassert_wait = 1;
309module_param(myri10ge_deassert_wait, int, S_IRUGO | S_IWUSR);
310MODULE_PARM_DESC(myri10ge_deassert_wait,
d1ce3a0f 311 "Wait when deasserting legacy interrupts");
0da34b6d
BG
312
313static int myri10ge_force_firmware = 0;
314module_param(myri10ge_force_firmware, int, S_IRUGO);
315MODULE_PARM_DESC(myri10ge_force_firmware,
d1ce3a0f 316 "Force firmware to assume aligned completions");
0da34b6d 317
0da34b6d
BG
318static int myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
319module_param(myri10ge_initial_mtu, int, S_IRUGO);
d1ce3a0f 320MODULE_PARM_DESC(myri10ge_initial_mtu, "Initial MTU");
0da34b6d
BG
321
322static int myri10ge_napi_weight = 64;
323module_param(myri10ge_napi_weight, int, S_IRUGO);
d1ce3a0f 324MODULE_PARM_DESC(myri10ge_napi_weight, "Set NAPI weight");
0da34b6d
BG
325
326static int myri10ge_watchdog_timeout = 1;
327module_param(myri10ge_watchdog_timeout, int, S_IRUGO);
d1ce3a0f 328MODULE_PARM_DESC(myri10ge_watchdog_timeout, "Set watchdog timeout");
0da34b6d
BG
329
330static int myri10ge_max_irq_loops = 1048576;
331module_param(myri10ge_max_irq_loops, int, S_IRUGO);
332MODULE_PARM_DESC(myri10ge_max_irq_loops,
d1ce3a0f 333 "Set stuck legacy IRQ detection threshold");
0da34b6d 334
c58ac5ca
BG
335#define MYRI10GE_MSG_DEFAULT NETIF_MSG_LINK
336
337static int myri10ge_debug = -1; /* defaults above */
338module_param(myri10ge_debug, int, 0);
339MODULE_PARM_DESC(myri10ge_debug, "Debug level (0=none,...,16=all)");
340
1e6e9342
AG
341static int myri10ge_lro_max_pkts = MYRI10GE_LRO_MAX_PKTS;
342module_param(myri10ge_lro_max_pkts, int, S_IRUGO);
d1ce3a0f
BG
343MODULE_PARM_DESC(myri10ge_lro_max_pkts,
344 "Number of LRO packets to be aggregated");
1e6e9342 345
dd50f336
BG
346static int myri10ge_fill_thresh = 256;
347module_param(myri10ge_fill_thresh, int, S_IRUGO | S_IWUSR);
d1ce3a0f 348MODULE_PARM_DESC(myri10ge_fill_thresh, "Number of empty rx slots allowed");
dd50f336 349
f181137f
BG
350static int myri10ge_reset_recover = 1;
351
0dcffac1
BG
352static int myri10ge_max_slices = 1;
353module_param(myri10ge_max_slices, int, S_IRUGO);
354MODULE_PARM_DESC(myri10ge_max_slices, "Max tx/rx queues");
355
4b860abf 356static int myri10ge_rss_hash = MXGEFW_RSS_HASH_TYPE_SRC_DST_PORT;
0dcffac1
BG
357module_param(myri10ge_rss_hash, int, S_IRUGO);
358MODULE_PARM_DESC(myri10ge_rss_hash, "Type of RSS hashing to do");
359
981813d8
BG
360static int myri10ge_dca = 1;
361module_param(myri10ge_dca, int, S_IRUGO);
362MODULE_PARM_DESC(myri10ge_dca, "Enable DCA if possible");
363
0da34b6d
BG
364#define MYRI10GE_FW_OFFSET 1024*1024
365#define MYRI10GE_HIGHPART_TO_U32(X) \
366(sizeof (X) == 8) ? ((u32)((u64)(X) >> 32)) : (0)
367#define MYRI10GE_LOWPART_TO_U32(X) ((u32)(X))
368
369#define myri10ge_pio_copy(to,from,size) __iowrite64_copy(to,from,size/8)
370
2f76216f 371static void myri10ge_set_multicast_list(struct net_device *dev);
61357325
SH
372static netdev_tx_t myri10ge_sw_tso(struct sk_buff *skb,
373 struct net_device *dev);
2f76216f 374
6250223e 375static inline void put_be32(__be32 val, __be32 __iomem * p)
40f6cff5 376{
6250223e 377 __raw_writel((__force __u32) val, (__force void __iomem *)p);
40f6cff5
AV
378}
379
59081825
BG
380static struct net_device_stats *myri10ge_get_stats(struct net_device *dev);
381
7d351035
RR
382static void set_fw_name(struct myri10ge_priv *mgp, char *name, bool allocated)
383{
384 if (mgp->fw_name_allocated)
385 kfree(mgp->fw_name);
386 mgp->fw_name = name;
387 mgp->fw_name_allocated = allocated;
388}
389
0da34b6d
BG
390static int
391myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd,
392 struct myri10ge_cmd *data, int atomic)
393{
394 struct mcp_cmd *buf;
395 char buf_bytes[sizeof(*buf) + 8];
396 struct mcp_cmd_response *response = mgp->cmd;
e700f9f4 397 char __iomem *cmd_addr = mgp->sram + MXGEFW_ETH_CMD;
0da34b6d
BG
398 u32 dma_low, dma_high, result, value;
399 int sleep_total = 0;
400
401 /* ensure buf is aligned to 8 bytes */
402 buf = (struct mcp_cmd *)ALIGN((unsigned long)buf_bytes, 8);
403
404 buf->data0 = htonl(data->data0);
405 buf->data1 = htonl(data->data1);
406 buf->data2 = htonl(data->data2);
407 buf->cmd = htonl(cmd);
408 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
409 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
410
411 buf->response_addr.low = htonl(dma_low);
412 buf->response_addr.high = htonl(dma_high);
40f6cff5 413 response->result = htonl(MYRI10GE_NO_RESPONSE_RESULT);
0da34b6d
BG
414 mb();
415 myri10ge_pio_copy(cmd_addr, buf, sizeof(*buf));
416
417 /* wait up to 15ms. Longest command is the DMA benchmark,
418 * which is capped at 5ms, but runs from a timeout handler
419 * that runs every 7.8ms. So a 15ms timeout leaves us with
420 * a 2.2ms margin
421 */
422 if (atomic) {
423 /* if atomic is set, do not sleep,
424 * and try to get the completion quickly
425 * (1ms will be enough for those commands) */
426 for (sleep_total = 0;
8e95a202
JP
427 sleep_total < 1000 &&
428 response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT);
bd2db0cf 429 sleep_total += 10) {
0da34b6d 430 udelay(10);
bd2db0cf
BG
431 mb();
432 }
0da34b6d
BG
433 } else {
434 /* use msleep for most command */
435 for (sleep_total = 0;
8e95a202
JP
436 sleep_total < 15 &&
437 response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT);
0da34b6d
BG
438 sleep_total++)
439 msleep(1);
440 }
441
442 result = ntohl(response->result);
443 value = ntohl(response->data);
444 if (result != MYRI10GE_NO_RESPONSE_RESULT) {
445 if (result == 0) {
446 data->data0 = value;
447 return 0;
85a7ea1b
BG
448 } else if (result == MXGEFW_CMD_UNKNOWN) {
449 return -ENOSYS;
5443e9ea
BG
450 } else if (result == MXGEFW_CMD_ERROR_UNALIGNED) {
451 return -E2BIG;
236bb5e6
BG
452 } else if (result == MXGEFW_CMD_ERROR_RANGE &&
453 cmd == MXGEFW_CMD_ENABLE_RSS_QUEUES &&
454 (data->
455 data1 & MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES) !=
456 0) {
457 return -ERANGE;
0da34b6d
BG
458 } else {
459 dev_err(&mgp->pdev->dev,
460 "command %d failed, result = %d\n",
461 cmd, result);
462 return -ENXIO;
463 }
464 }
465
466 dev_err(&mgp->pdev->dev, "command %d timed out, result = %d\n",
467 cmd, result);
468 return -EAGAIN;
469}
470
471/*
472 * The eeprom strings on the lanaiX have the format
473 * SN=x\0
474 * MAC=x:x:x:x:x:x\0
475 * PT:ddd mmm xx xx:xx:xx xx\0
476 * PV:ddd mmm xx xx:xx:xx xx\0
477 */
478static int myri10ge_read_mac_addr(struct myri10ge_priv *mgp)
479{
480 char *ptr, *limit;
481 int i;
482
483 ptr = mgp->eeprom_strings;
484 limit = mgp->eeprom_strings + MYRI10GE_EEPROM_STRINGS_SIZE;
485
486 while (*ptr != '\0' && ptr < limit) {
487 if (memcmp(ptr, "MAC=", 4) == 0) {
488 ptr += 4;
489 mgp->mac_addr_string = ptr;
490 for (i = 0; i < 6; i++) {
491 if ((ptr + 2) > limit)
492 goto abort;
493 mgp->mac_addr[i] =
494 simple_strtoul(ptr, &ptr, 16);
495 ptr += 1;
496 }
497 }
c0bf8801
BG
498 if (memcmp(ptr, "PC=", 3) == 0) {
499 ptr += 3;
500 mgp->product_code_string = ptr;
501 }
0da34b6d
BG
502 if (memcmp((const void *)ptr, "SN=", 3) == 0) {
503 ptr += 3;
504 mgp->serial_number = simple_strtoul(ptr, &ptr, 10);
505 }
506 while (ptr < limit && *ptr++) ;
507 }
508
509 return 0;
510
511abort:
512 dev_err(&mgp->pdev->dev, "failed to parse eeprom_strings\n");
513 return -ENXIO;
514}
515
516/*
517 * Enable or disable periodic RDMAs from the host to make certain
518 * chipsets resend dropped PCIe messages
519 */
520
521static void myri10ge_dummy_rdma(struct myri10ge_priv *mgp, int enable)
522{
523 char __iomem *submit;
f8fd57c1 524 __be32 buf[16] __attribute__ ((__aligned__(8)));
0da34b6d
BG
525 u32 dma_low, dma_high;
526 int i;
527
528 /* clear confirmation addr */
529 mgp->cmd->data = 0;
530 mb();
531
532 /* send a rdma command to the PCIe engine, and wait for the
533 * response in the confirmation address. The firmware should
534 * write a -1 there to indicate it is alive and well
535 */
536 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
537 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
538
539 buf[0] = htonl(dma_high); /* confirm addr MSW */
540 buf[1] = htonl(dma_low); /* confirm addr LSW */
40f6cff5 541 buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */
0da34b6d
BG
542 buf[3] = htonl(dma_high); /* dummy addr MSW */
543 buf[4] = htonl(dma_low); /* dummy addr LSW */
544 buf[5] = htonl(enable); /* enable? */
545
e700f9f4 546 submit = mgp->sram + MXGEFW_BOOT_DUMMY_RDMA;
0da34b6d
BG
547
548 myri10ge_pio_copy(submit, &buf, sizeof(buf));
549 for (i = 0; mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA && i < 20; i++)
550 msleep(1);
551 if (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA)
552 dev_err(&mgp->pdev->dev, "dummy rdma %s failed\n",
553 (enable ? "enable" : "disable"));
554}
555
556static int
557myri10ge_validate_firmware(struct myri10ge_priv *mgp,
558 struct mcp_gen_header *hdr)
559{
560 struct device *dev = &mgp->pdev->dev;
0da34b6d
BG
561
562 /* check firmware type */
563 if (ntohl(hdr->mcp_type) != MCP_TYPE_ETH) {
564 dev_err(dev, "Bad firmware type: 0x%x\n", ntohl(hdr->mcp_type));
565 return -EINVAL;
566 }
567
568 /* save firmware version for ethtool */
569 strncpy(mgp->fw_version, hdr->version, sizeof(mgp->fw_version));
570
9dc6f0e7
BG
571 sscanf(mgp->fw_version, "%d.%d.%d", &mgp->fw_ver_major,
572 &mgp->fw_ver_minor, &mgp->fw_ver_tiny);
0da34b6d 573
8e95a202
JP
574 if (!(mgp->fw_ver_major == MXGEFW_VERSION_MAJOR &&
575 mgp->fw_ver_minor == MXGEFW_VERSION_MINOR)) {
0da34b6d
BG
576 dev_err(dev, "Found firmware version %s\n", mgp->fw_version);
577 dev_err(dev, "Driver needs %d.%d\n", MXGEFW_VERSION_MAJOR,
578 MXGEFW_VERSION_MINOR);
579 return -EINVAL;
580 }
581 return 0;
582}
583
584static int myri10ge_load_hotplug_firmware(struct myri10ge_priv *mgp, u32 * size)
585{
586 unsigned crc, reread_crc;
587 const struct firmware *fw;
588 struct device *dev = &mgp->pdev->dev;
b0d31d6b 589 unsigned char *fw_readback;
0da34b6d
BG
590 struct mcp_gen_header *hdr;
591 size_t hdr_offset;
592 int status;
e454358a 593 unsigned i;
0da34b6d
BG
594
595 if ((status = request_firmware(&fw, mgp->fw_name, dev)) < 0) {
596 dev_err(dev, "Unable to load %s firmware image via hotplug\n",
597 mgp->fw_name);
598 status = -EINVAL;
599 goto abort_with_nothing;
600 }
601
602 /* check size */
603
604 if (fw->size >= mgp->sram_size - MYRI10GE_FW_OFFSET ||
605 fw->size < MCP_HEADER_PTR_OFFSET + 4) {
606 dev_err(dev, "Firmware size invalid:%d\n", (int)fw->size);
607 status = -EINVAL;
608 goto abort_with_fw;
609 }
610
611 /* check id */
40f6cff5 612 hdr_offset = ntohl(*(__be32 *) (fw->data + MCP_HEADER_PTR_OFFSET));
0da34b6d
BG
613 if ((hdr_offset & 3) || hdr_offset + sizeof(*hdr) > fw->size) {
614 dev_err(dev, "Bad firmware file\n");
615 status = -EINVAL;
616 goto abort_with_fw;
617 }
618 hdr = (void *)(fw->data + hdr_offset);
619
620 status = myri10ge_validate_firmware(mgp, hdr);
621 if (status != 0)
622 goto abort_with_fw;
623
624 crc = crc32(~0, fw->data, fw->size);
e454358a
BG
625 for (i = 0; i < fw->size; i += 256) {
626 myri10ge_pio_copy(mgp->sram + MYRI10GE_FW_OFFSET + i,
627 fw->data + i,
628 min(256U, (unsigned)(fw->size - i)));
629 mb();
630 readb(mgp->sram);
b10c0668 631 }
b0d31d6b
DW
632 fw_readback = vmalloc(fw->size);
633 if (!fw_readback) {
634 status = -ENOMEM;
635 goto abort_with_fw;
636 }
0da34b6d 637 /* corruption checking is good for parity recovery and buggy chipset */
b0d31d6b
DW
638 memcpy_fromio(fw_readback, mgp->sram + MYRI10GE_FW_OFFSET, fw->size);
639 reread_crc = crc32(~0, fw_readback, fw->size);
640 vfree(fw_readback);
0da34b6d
BG
641 if (crc != reread_crc) {
642 dev_err(dev, "CRC failed(fw-len=%u), got 0x%x (expect 0x%x)\n",
643 (unsigned)fw->size, reread_crc, crc);
644 status = -EIO;
645 goto abort_with_fw;
646 }
647 *size = (u32) fw->size;
648
649abort_with_fw:
650 release_firmware(fw);
651
652abort_with_nothing:
653 return status;
654}
655
656static int myri10ge_adopt_running_firmware(struct myri10ge_priv *mgp)
657{
658 struct mcp_gen_header *hdr;
659 struct device *dev = &mgp->pdev->dev;
660 const size_t bytes = sizeof(struct mcp_gen_header);
661 size_t hdr_offset;
662 int status;
663
664 /* find running firmware header */
66341fff 665 hdr_offset = swab32(readl(mgp->sram + MCP_HEADER_PTR_OFFSET));
0da34b6d
BG
666
667 if ((hdr_offset & 3) || hdr_offset + sizeof(*hdr) > mgp->sram_size) {
668 dev_err(dev, "Running firmware has bad header offset (%d)\n",
669 (int)hdr_offset);
670 return -EIO;
671 }
672
673 /* copy header of running firmware from SRAM to host memory to
674 * validate firmware */
675 hdr = kmalloc(bytes, GFP_KERNEL);
676 if (hdr == NULL) {
677 dev_err(dev, "could not malloc firmware hdr\n");
678 return -ENOMEM;
679 }
680 memcpy_fromio(hdr, mgp->sram + hdr_offset, bytes);
681 status = myri10ge_validate_firmware(mgp, hdr);
682 kfree(hdr);
9dc6f0e7
BG
683
684 /* check to see if adopted firmware has bug where adopting
685 * it will cause broadcasts to be filtered unless the NIC
686 * is kept in ALLMULTI mode */
687 if (mgp->fw_ver_major == 1 && mgp->fw_ver_minor == 4 &&
688 mgp->fw_ver_tiny >= 4 && mgp->fw_ver_tiny <= 11) {
689 mgp->adopted_rx_filter_bug = 1;
690 dev_warn(dev, "Adopting fw %d.%d.%d: "
691 "working around rx filter bug\n",
692 mgp->fw_ver_major, mgp->fw_ver_minor,
693 mgp->fw_ver_tiny);
694 }
0da34b6d
BG
695 return status;
696}
697
0178ec3d 698static int myri10ge_get_firmware_capabilities(struct myri10ge_priv *mgp)
fa0a90d9
BG
699{
700 struct myri10ge_cmd cmd;
701 int status;
702
703 /* probe for IPv6 TSO support */
704 mgp->features = NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_TSO;
705 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_TSO6_HDR_SIZE,
706 &cmd, 0);
707 if (status == 0) {
708 mgp->max_tso6 = cmd.data0;
709 mgp->features |= NETIF_F_TSO6;
710 }
711
712 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_RX_RING_SIZE, &cmd, 0);
713 if (status != 0) {
714 dev_err(&mgp->pdev->dev,
715 "failed MXGEFW_CMD_GET_RX_RING_SIZE\n");
716 return -ENXIO;
717 }
718
719 mgp->max_intr_slots = 2 * (cmd.data0 / sizeof(struct mcp_dma_addr));
720
721 return 0;
722}
723
0dcffac1 724static int myri10ge_load_firmware(struct myri10ge_priv *mgp, int adopt)
0da34b6d
BG
725{
726 char __iomem *submit;
f8fd57c1 727 __be32 buf[16] __attribute__ ((__aligned__(8)));
0da34b6d
BG
728 u32 dma_low, dma_high, size;
729 int status, i;
730
b10c0668 731 size = 0;
0da34b6d
BG
732 status = myri10ge_load_hotplug_firmware(mgp, &size);
733 if (status) {
0dcffac1
BG
734 if (!adopt)
735 return status;
0da34b6d
BG
736 dev_warn(&mgp->pdev->dev, "hotplug firmware loading failed\n");
737
738 /* Do not attempt to adopt firmware if there
739 * was a bad crc */
740 if (status == -EIO)
741 return status;
742
743 status = myri10ge_adopt_running_firmware(mgp);
744 if (status != 0) {
745 dev_err(&mgp->pdev->dev,
746 "failed to adopt running firmware\n");
747 return status;
748 }
749 dev_info(&mgp->pdev->dev,
750 "Successfully adopted running firmware\n");
b53bef84 751 if (mgp->tx_boundary == 4096) {
0da34b6d
BG
752 dev_warn(&mgp->pdev->dev,
753 "Using firmware currently running on NIC"
754 ". For optimal\n");
755 dev_warn(&mgp->pdev->dev,
756 "performance consider loading optimized "
757 "firmware\n");
758 dev_warn(&mgp->pdev->dev, "via hotplug\n");
759 }
760
7d351035 761 set_fw_name(mgp, "adopted", false);
b53bef84 762 mgp->tx_boundary = 2048;
fa0a90d9
BG
763 myri10ge_dummy_rdma(mgp, 1);
764 status = myri10ge_get_firmware_capabilities(mgp);
0da34b6d
BG
765 return status;
766 }
767
768 /* clear confirmation addr */
769 mgp->cmd->data = 0;
770 mb();
771
772 /* send a reload command to the bootstrap MCP, and wait for the
773 * response in the confirmation address. The firmware should
774 * write a -1 there to indicate it is alive and well
775 */
776 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
777 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
778
779 buf[0] = htonl(dma_high); /* confirm addr MSW */
780 buf[1] = htonl(dma_low); /* confirm addr LSW */
40f6cff5 781 buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */
0da34b6d
BG
782
783 /* FIX: All newest firmware should un-protect the bottom of
784 * the sram before handoff. However, the very first interfaces
785 * do not. Therefore the handoff copy must skip the first 8 bytes
786 */
787 buf[3] = htonl(MYRI10GE_FW_OFFSET + 8); /* where the code starts */
788 buf[4] = htonl(size - 8); /* length of code */
789 buf[5] = htonl(8); /* where to copy to */
790 buf[6] = htonl(0); /* where to jump to */
791
e700f9f4 792 submit = mgp->sram + MXGEFW_BOOT_HANDOFF;
0da34b6d
BG
793
794 myri10ge_pio_copy(submit, &buf, sizeof(buf));
795 mb();
796 msleep(1);
797 mb();
798 i = 0;
d93ca2a4
BG
799 while (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA && i < 9) {
800 msleep(1 << i);
0da34b6d
BG
801 i++;
802 }
803 if (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA) {
804 dev_err(&mgp->pdev->dev, "handoff failed\n");
805 return -ENXIO;
806 }
9a71db72 807 myri10ge_dummy_rdma(mgp, 1);
fa0a90d9 808 status = myri10ge_get_firmware_capabilities(mgp);
0da34b6d 809
fa0a90d9 810 return status;
0da34b6d
BG
811}
812
813static int myri10ge_update_mac_address(struct myri10ge_priv *mgp, u8 * addr)
814{
815 struct myri10ge_cmd cmd;
816 int status;
817
818 cmd.data0 = ((addr[0] << 24) | (addr[1] << 16)
819 | (addr[2] << 8) | addr[3]);
820
821 cmd.data1 = ((addr[4] << 8) | (addr[5]));
822
823 status = myri10ge_send_cmd(mgp, MXGEFW_SET_MAC_ADDRESS, &cmd, 0);
824 return status;
825}
826
827static int myri10ge_change_pause(struct myri10ge_priv *mgp, int pause)
828{
829 struct myri10ge_cmd cmd;
830 int status, ctl;
831
832 ctl = pause ? MXGEFW_ENABLE_FLOW_CONTROL : MXGEFW_DISABLE_FLOW_CONTROL;
833 status = myri10ge_send_cmd(mgp, ctl, &cmd, 0);
834
835 if (status) {
78ca90ea 836 netdev_err(mgp->dev, "Failed to set flow control mode\n");
0da34b6d
BG
837 return status;
838 }
839 mgp->pause = pause;
840 return 0;
841}
842
843static void
844myri10ge_change_promisc(struct myri10ge_priv *mgp, int promisc, int atomic)
845{
846 struct myri10ge_cmd cmd;
847 int status, ctl;
848
849 ctl = promisc ? MXGEFW_ENABLE_PROMISC : MXGEFW_DISABLE_PROMISC;
850 status = myri10ge_send_cmd(mgp, ctl, &cmd, atomic);
851 if (status)
78ca90ea 852 netdev_err(mgp->dev, "Failed to set promisc mode\n");
0da34b6d
BG
853}
854
0d6ac257 855static int myri10ge_dma_test(struct myri10ge_priv *mgp, int test_type)
0da34b6d
BG
856{
857 struct myri10ge_cmd cmd;
858 int status;
0da34b6d 859 u32 len;
34fdccea
BG
860 struct page *dmatest_page;
861 dma_addr_t dmatest_bus;
0d6ac257
BG
862 char *test = " ";
863
864 dmatest_page = alloc_page(GFP_KERNEL);
865 if (!dmatest_page)
866 return -ENOMEM;
867 dmatest_bus = pci_map_page(mgp->pdev, dmatest_page, 0, PAGE_SIZE,
868 DMA_BIDIRECTIONAL);
869
870 /* Run a small DMA test.
871 * The magic multipliers to the length tell the firmware
872 * to do DMA read, write, or read+write tests. The
873 * results are returned in cmd.data0. The upper 16
874 * bits or the return is the number of transfers completed.
875 * The lower 16 bits is the time in 0.5us ticks that the
876 * transfers took to complete.
877 */
878
b53bef84 879 len = mgp->tx_boundary;
0d6ac257
BG
880
881 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
882 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
883 cmd.data2 = len * 0x10000;
884 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
885 if (status != 0) {
886 test = "read";
887 goto abort;
888 }
889 mgp->read_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
890 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
891 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
892 cmd.data2 = len * 0x1;
893 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
894 if (status != 0) {
895 test = "write";
896 goto abort;
897 }
898 mgp->write_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
899
900 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
901 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
902 cmd.data2 = len * 0x10001;
903 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
904 if (status != 0) {
905 test = "read/write";
906 goto abort;
907 }
908 mgp->read_write_dma = ((cmd.data0 >> 16) * len * 2 * 2) /
909 (cmd.data0 & 0xffff);
910
911abort:
912 pci_unmap_page(mgp->pdev, dmatest_bus, PAGE_SIZE, DMA_BIDIRECTIONAL);
913 put_page(dmatest_page);
914
915 if (status != 0 && test_type != MXGEFW_CMD_UNALIGNED_TEST)
916 dev_warn(&mgp->pdev->dev, "DMA %s benchmark failed: %d\n",
917 test, status);
918
919 return status;
920}
921
922static int myri10ge_reset(struct myri10ge_priv *mgp)
923{
924 struct myri10ge_cmd cmd;
0dcffac1
BG
925 struct myri10ge_slice_state *ss;
926 int i, status;
0d6ac257 927 size_t bytes;
5dd2d332 928#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
929 unsigned long dca_tag_off;
930#endif
0da34b6d
BG
931
932 /* try to send a reset command to the card to see if it
933 * is alive */
934 memset(&cmd, 0, sizeof(cmd));
935 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_RESET, &cmd, 0);
936 if (status != 0) {
937 dev_err(&mgp->pdev->dev, "failed reset\n");
938 return -ENXIO;
939 }
0d6ac257
BG
940
941 (void)myri10ge_dma_test(mgp, MXGEFW_DMA_TEST);
0dcffac1
BG
942 /*
943 * Use non-ndis mcp_slot (eg, 4 bytes total,
944 * no toeplitz hash value returned. Older firmware will
945 * not understand this command, but will use the correct
946 * sized mcp_slot, so we ignore error returns
947 */
948 cmd.data0 = MXGEFW_RSS_MCP_SLOT_TYPE_MIN;
949 (void)myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_MCP_SLOT_TYPE, &cmd, 0);
0da34b6d
BG
950
951 /* Now exchange information about interrupts */
952
0dcffac1 953 bytes = mgp->max_intr_slots * sizeof(*mgp->ss[0].rx_done.entry);
0da34b6d
BG
954 cmd.data0 = (u32) bytes;
955 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_SIZE, &cmd, 0);
0dcffac1
BG
956
957 /*
958 * Even though we already know how many slices are supported
959 * via myri10ge_probe_slices() MXGEFW_CMD_GET_MAX_RSS_QUEUES
960 * has magic side effects, and must be called after a reset.
961 * It must be called prior to calling any RSS related cmds,
962 * including assigning an interrupt queue for anything but
963 * slice 0. It must also be called *after*
964 * MXGEFW_CMD_SET_INTRQ_SIZE, since the intrq size is used by
965 * the firmware to compute offsets.
966 */
967
968 if (mgp->num_slices > 1) {
969
970 /* ask the maximum number of slices it supports */
971 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_RSS_QUEUES,
972 &cmd, 0);
973 if (status != 0) {
974 dev_err(&mgp->pdev->dev,
975 "failed to get number of slices\n");
976 }
977
978 /*
979 * MXGEFW_CMD_ENABLE_RSS_QUEUES must be called prior
980 * to setting up the interrupt queue DMA
981 */
982
983 cmd.data0 = mgp->num_slices;
236bb5e6
BG
984 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
985 if (mgp->dev->real_num_tx_queues > 1)
986 cmd.data1 |= MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES;
0dcffac1
BG
987 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ENABLE_RSS_QUEUES,
988 &cmd, 0);
236bb5e6
BG
989
990 /* Firmware older than 1.4.32 only supports multiple
991 * RX queues, so if we get an error, first retry using a
992 * single TX queue before giving up */
993 if (status != 0 && mgp->dev->real_num_tx_queues > 1) {
c9920268 994 netif_set_real_num_tx_queues(mgp->dev, 1);
236bb5e6
BG
995 cmd.data0 = mgp->num_slices;
996 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
997 status = myri10ge_send_cmd(mgp,
998 MXGEFW_CMD_ENABLE_RSS_QUEUES,
999 &cmd, 0);
1000 }
1001
0dcffac1
BG
1002 if (status != 0) {
1003 dev_err(&mgp->pdev->dev,
1004 "failed to set number of slices\n");
1005
1006 return status;
1007 }
1008 }
1009 for (i = 0; i < mgp->num_slices; i++) {
1010 ss = &mgp->ss[i];
1011 cmd.data0 = MYRI10GE_LOWPART_TO_U32(ss->rx_done.bus);
1012 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(ss->rx_done.bus);
1013 cmd.data2 = i;
1014 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_DMA,
1015 &cmd, 0);
1016 };
0da34b6d
BG
1017
1018 status |=
1019 myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_ACK_OFFSET, &cmd, 0);
0dcffac1
BG
1020 for (i = 0; i < mgp->num_slices; i++) {
1021 ss = &mgp->ss[i];
1022 ss->irq_claim =
1023 (__iomem __be32 *) (mgp->sram + cmd.data0 + 8 * i);
1024 }
df30a740
BG
1025 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_DEASSERT_OFFSET,
1026 &cmd, 0);
1027 mgp->irq_deassert = (__iomem __be32 *) (mgp->sram + cmd.data0);
0da34b6d 1028
0da34b6d
BG
1029 status |= myri10ge_send_cmd
1030 (mgp, MXGEFW_CMD_GET_INTR_COAL_DELAY_OFFSET, &cmd, 0);
40f6cff5 1031 mgp->intr_coal_delay_ptr = (__iomem __be32 *) (mgp->sram + cmd.data0);
0da34b6d
BG
1032 if (status != 0) {
1033 dev_err(&mgp->pdev->dev, "failed set interrupt parameters\n");
1034 return status;
1035 }
40f6cff5 1036 put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr);
0da34b6d 1037
5dd2d332 1038#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
1039 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_DCA_OFFSET, &cmd, 0);
1040 dca_tag_off = cmd.data0;
1041 for (i = 0; i < mgp->num_slices; i++) {
1042 ss = &mgp->ss[i];
1043 if (status == 0) {
1044 ss->dca_tag = (__iomem __be32 *)
1045 (mgp->sram + dca_tag_off + 4 * i);
1046 } else {
1047 ss->dca_tag = NULL;
1048 }
1049 }
4ee2ac51 1050#endif /* CONFIG_MYRI10GE_DCA */
981813d8 1051
0da34b6d 1052 /* reset mcp/driver shared state back to 0 */
0dcffac1 1053
c58ac5ca 1054 mgp->link_changes = 0;
0dcffac1
BG
1055 for (i = 0; i < mgp->num_slices; i++) {
1056 ss = &mgp->ss[i];
1057
1058 memset(ss->rx_done.entry, 0, bytes);
1059 ss->tx.req = 0;
1060 ss->tx.done = 0;
1061 ss->tx.pkt_start = 0;
1062 ss->tx.pkt_done = 0;
1063 ss->rx_big.cnt = 0;
1064 ss->rx_small.cnt = 0;
1065 ss->rx_done.idx = 0;
1066 ss->rx_done.cnt = 0;
1067 ss->tx.wake_queue = 0;
1068 ss->tx.stop_queue = 0;
1069 }
1070
0da34b6d 1071 status = myri10ge_update_mac_address(mgp, mgp->dev->dev_addr);
0da34b6d 1072 myri10ge_change_pause(mgp, mgp->pause);
2f76216f 1073 myri10ge_set_multicast_list(mgp->dev);
0da34b6d
BG
1074 return status;
1075}
1076
5dd2d332 1077#ifdef CONFIG_MYRI10GE_DCA
ef09aadf
AG
1078static int myri10ge_toggle_relaxed(struct pci_dev *pdev, int on)
1079{
1080 int ret, cap, err;
1081 u16 ctl;
1082
1083 cap = pci_find_capability(pdev, PCI_CAP_ID_EXP);
1084 if (!cap)
1085 return 0;
1086
1087 err = pci_read_config_word(pdev, cap + PCI_EXP_DEVCTL, &ctl);
1088 ret = (ctl & PCI_EXP_DEVCTL_RELAX_EN) >> 4;
1089 if (ret != on) {
1090 ctl &= ~PCI_EXP_DEVCTL_RELAX_EN;
1091 ctl |= (on << 4);
1092 pci_write_config_word(pdev, cap + PCI_EXP_DEVCTL, ctl);
1093 }
1094 return ret;
1095}
1096
981813d8
BG
1097static void
1098myri10ge_write_dca(struct myri10ge_slice_state *ss, int cpu, int tag)
1099{
981813d8
BG
1100 ss->cached_dca_tag = tag;
1101 put_be32(htonl(tag), ss->dca_tag);
1102}
1103
1104static inline void myri10ge_update_dca(struct myri10ge_slice_state *ss)
1105{
1106 int cpu = get_cpu();
1107 int tag;
1108
1109 if (cpu != ss->cpu) {
ef09aadf 1110 tag = dca3_get_tag(&ss->mgp->pdev->dev, cpu);
981813d8
BG
1111 if (ss->cached_dca_tag != tag)
1112 myri10ge_write_dca(ss, cpu, tag);
ef09aadf 1113 ss->cpu = cpu;
981813d8
BG
1114 }
1115 put_cpu();
1116}
1117
1118static void myri10ge_setup_dca(struct myri10ge_priv *mgp)
1119{
1120 int err, i;
1121 struct pci_dev *pdev = mgp->pdev;
1122
1123 if (mgp->ss[0].dca_tag == NULL || mgp->dca_enabled)
1124 return;
1125 if (!myri10ge_dca) {
1126 dev_err(&pdev->dev, "dca disabled by administrator\n");
1127 return;
1128 }
1129 err = dca_add_requester(&pdev->dev);
1130 if (err) {
330554cb
BG
1131 if (err != -ENODEV)
1132 dev_err(&pdev->dev,
1133 "dca_add_requester() failed, err=%d\n", err);
981813d8
BG
1134 return;
1135 }
ef09aadf 1136 mgp->relaxed_order = myri10ge_toggle_relaxed(pdev, 0);
981813d8 1137 mgp->dca_enabled = 1;
ef09aadf
AG
1138 for (i = 0; i < mgp->num_slices; i++) {
1139 mgp->ss[i].cpu = -1;
1140 mgp->ss[i].cached_dca_tag = -1;
1141 myri10ge_update_dca(&mgp->ss[i]);
1142 }
981813d8
BG
1143}
1144
1145static void myri10ge_teardown_dca(struct myri10ge_priv *mgp)
1146{
1147 struct pci_dev *pdev = mgp->pdev;
1148 int err;
1149
1150 if (!mgp->dca_enabled)
1151 return;
1152 mgp->dca_enabled = 0;
ef09aadf
AG
1153 if (mgp->relaxed_order)
1154 myri10ge_toggle_relaxed(pdev, 1);
981813d8
BG
1155 err = dca_remove_requester(&pdev->dev);
1156}
1157
1158static int myri10ge_notify_dca_device(struct device *dev, void *data)
1159{
1160 struct myri10ge_priv *mgp;
1161 unsigned long event;
1162
1163 mgp = dev_get_drvdata(dev);
1164 event = *(unsigned long *)data;
1165
1166 if (event == DCA_PROVIDER_ADD)
1167 myri10ge_setup_dca(mgp);
1168 else if (event == DCA_PROVIDER_REMOVE)
1169 myri10ge_teardown_dca(mgp);
1170 return 0;
1171}
4ee2ac51 1172#endif /* CONFIG_MYRI10GE_DCA */
981813d8 1173
0da34b6d
BG
1174static inline void
1175myri10ge_submit_8rx(struct mcp_kreq_ether_recv __iomem * dst,
1176 struct mcp_kreq_ether_recv *src)
1177{
40f6cff5 1178 __be32 low;
0da34b6d
BG
1179
1180 low = src->addr_low;
284901a9 1181 src->addr_low = htonl(DMA_BIT_MASK(32));
e67bda55
BG
1182 myri10ge_pio_copy(dst, src, 4 * sizeof(*src));
1183 mb();
1184 myri10ge_pio_copy(dst + 4, src + 4, 4 * sizeof(*src));
0da34b6d
BG
1185 mb();
1186 src->addr_low = low;
40f6cff5 1187 put_be32(low, &dst->addr_low);
0da34b6d
BG
1188 mb();
1189}
1190
40f6cff5 1191static inline void myri10ge_vlan_ip_csum(struct sk_buff *skb, __wsum hw_csum)
0da34b6d
BG
1192{
1193 struct vlan_hdr *vh = (struct vlan_hdr *)(skb->data);
1194
40f6cff5 1195 if ((skb->protocol == htons(ETH_P_8021Q)) &&
0da34b6d
BG
1196 (vh->h_vlan_encapsulated_proto == htons(ETH_P_IP) ||
1197 vh->h_vlan_encapsulated_proto == htons(ETH_P_IPV6))) {
1198 skb->csum = hw_csum;
84fa7933 1199 skb->ip_summed = CHECKSUM_COMPLETE;
0da34b6d
BG
1200 }
1201}
1202
dd50f336
BG
1203static inline void
1204myri10ge_rx_skb_build(struct sk_buff *skb, u8 * va,
1205 struct skb_frag_struct *rx_frags, int len, int hlen)
1206{
1207 struct skb_frag_struct *skb_frags;
1208
1209 skb->len = skb->data_len = len;
1210 skb->truesize = len + sizeof(struct sk_buff);
1211 /* attach the page(s) */
1212
1213 skb_frags = skb_shinfo(skb)->frags;
1214 while (len > 0) {
1215 memcpy(skb_frags, rx_frags, sizeof(*skb_frags));
1216 len -= rx_frags->size;
1217 skb_frags++;
1218 rx_frags++;
1219 skb_shinfo(skb)->nr_frags++;
1220 }
1221
1222 /* pskb_may_pull is not available in irq context, but
1223 * skb_pull() (for ether_pad and eth_type_trans()) requires
1224 * the beginning of the packet in skb_headlen(), move it
1225 * manually */
27d7ff46 1226 skb_copy_to_linear_data(skb, va, hlen);
dd50f336
BG
1227 skb_shinfo(skb)->frags[0].page_offset += hlen;
1228 skb_shinfo(skb)->frags[0].size -= hlen;
1229 skb->data_len -= hlen;
1230 skb->tail += hlen;
1231 skb_pull(skb, MXGEFW_PAD);
1232}
1233
1234static void
1235myri10ge_alloc_rx_pages(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx,
1236 int bytes, int watchdog)
1237{
1238 struct page *page;
1239 int idx;
2a3f2790
BG
1240#if MYRI10GE_ALLOC_SIZE > 4096
1241 int end_offset;
1242#endif
dd50f336
BG
1243
1244 if (unlikely(rx->watchdog_needed && !watchdog))
1245 return;
1246
1247 /* try to refill entire ring */
1248 while (rx->fill_cnt != (rx->cnt + rx->mask + 1)) {
1249 idx = rx->fill_cnt & rx->mask;
ae8509b1 1250 if (rx->page_offset + bytes <= MYRI10GE_ALLOC_SIZE) {
dd50f336
BG
1251 /* we can use part of previous page */
1252 get_page(rx->page);
1253 } else {
1254 /* we need a new page */
1255 page =
1256 alloc_pages(GFP_ATOMIC | __GFP_COMP,
1257 MYRI10GE_ALLOC_ORDER);
1258 if (unlikely(page == NULL)) {
1259 if (rx->fill_cnt - rx->cnt < 16)
1260 rx->watchdog_needed = 1;
1261 return;
1262 }
1263 rx->page = page;
1264 rx->page_offset = 0;
1265 rx->bus = pci_map_page(mgp->pdev, page, 0,
1266 MYRI10GE_ALLOC_SIZE,
1267 PCI_DMA_FROMDEVICE);
1268 }
1269 rx->info[idx].page = rx->page;
1270 rx->info[idx].page_offset = rx->page_offset;
1271 /* note that this is the address of the start of the
1272 * page */
c755b4b6 1273 dma_unmap_addr_set(&rx->info[idx], bus, rx->bus);
dd50f336
BG
1274 rx->shadow[idx].addr_low =
1275 htonl(MYRI10GE_LOWPART_TO_U32(rx->bus) + rx->page_offset);
1276 rx->shadow[idx].addr_high =
1277 htonl(MYRI10GE_HIGHPART_TO_U32(rx->bus));
1278
1279 /* start next packet on a cacheline boundary */
1280 rx->page_offset += SKB_DATA_ALIGN(bytes);
ae8509b1
BG
1281
1282#if MYRI10GE_ALLOC_SIZE > 4096
1283 /* don't cross a 4KB boundary */
2a3f2790
BG
1284 end_offset = rx->page_offset + bytes - 1;
1285 if ((unsigned)(rx->page_offset ^ end_offset) > 4095)
1286 rx->page_offset = end_offset & ~4095;
ae8509b1 1287#endif
dd50f336
BG
1288 rx->fill_cnt++;
1289
1290 /* copy 8 descriptors to the firmware at a time */
1291 if ((idx & 7) == 7) {
e454e7e2
BG
1292 myri10ge_submit_8rx(&rx->lanai[idx - 7],
1293 &rx->shadow[idx - 7]);
dd50f336
BG
1294 }
1295 }
1296}
1297
1298static inline void
1299myri10ge_unmap_rx_page(struct pci_dev *pdev,
1300 struct myri10ge_rx_buffer_state *info, int bytes)
1301{
1302 /* unmap the recvd page if we're the only or last user of it */
1303 if (bytes >= MYRI10GE_ALLOC_SIZE / 2 ||
1304 (info->page_offset + 2 * bytes) > MYRI10GE_ALLOC_SIZE) {
c755b4b6 1305 pci_unmap_page(pdev, (dma_unmap_addr(info, bus)
dd50f336
BG
1306 & ~(MYRI10GE_ALLOC_SIZE - 1)),
1307 MYRI10GE_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
1308 }
1309}
1310
1311#define MYRI10GE_HLEN 64 /* The number of bytes to copy from a
1312 * page into an skb */
1313
1314static inline int
b3cd9657
SG
1315myri10ge_rx_done(struct myri10ge_slice_state *ss, int len, __wsum csum,
1316 int lro_enabled)
dd50f336 1317{
b53bef84 1318 struct myri10ge_priv *mgp = ss->mgp;
dd50f336
BG
1319 struct sk_buff *skb;
1320 struct skb_frag_struct rx_frags[MYRI10GE_MAX_FRAGS_PER_FRAME];
b3cd9657
SG
1321 struct myri10ge_rx_buf *rx;
1322 int i, idx, hlen, remainder, bytes;
dd50f336
BG
1323 struct pci_dev *pdev = mgp->pdev;
1324 struct net_device *dev = mgp->dev;
1325 u8 *va;
1326
b3cd9657
SG
1327 if (len <= mgp->small_bytes) {
1328 rx = &ss->rx_small;
1329 bytes = mgp->small_bytes;
1330 } else {
1331 rx = &ss->rx_big;
1332 bytes = mgp->big_bytes;
1333 }
1334
dd50f336
BG
1335 len += MXGEFW_PAD;
1336 idx = rx->cnt & rx->mask;
1337 va = page_address(rx->info[idx].page) + rx->info[idx].page_offset;
1338 prefetch(va);
1339 /* Fill skb_frag_struct(s) with data from our receive */
1340 for (i = 0, remainder = len; remainder > 0; i++) {
1341 myri10ge_unmap_rx_page(pdev, &rx->info[idx], bytes);
1342 rx_frags[i].page = rx->info[idx].page;
1343 rx_frags[i].page_offset = rx->info[idx].page_offset;
1344 if (remainder < MYRI10GE_ALLOC_SIZE)
1345 rx_frags[i].size = remainder;
1346 else
1347 rx_frags[i].size = MYRI10GE_ALLOC_SIZE;
1348 rx->cnt++;
1349 idx = rx->cnt & rx->mask;
1350 remainder -= MYRI10GE_ALLOC_SIZE;
1351 }
1352
b3cd9657 1353 if (lro_enabled) {
1e6e9342
AG
1354 rx_frags[0].page_offset += MXGEFW_PAD;
1355 rx_frags[0].size -= MXGEFW_PAD;
1356 len -= MXGEFW_PAD;
b53bef84 1357 lro_receive_frags(&ss->rx_done.lro_mgr, rx_frags,
b53bef84 1358 /* opaque, will come back in get_frag_header */
0dcffac1 1359 len, len,
b53bef84 1360 (void *)(__force unsigned long)csum, csum);
0dcffac1 1361
1e6e9342
AG
1362 return 1;
1363 }
1364
dd50f336
BG
1365 hlen = MYRI10GE_HLEN > len ? len : MYRI10GE_HLEN;
1366
e636b2ea
BG
1367 /* allocate an skb to attach the page(s) to. This is done
1368 * after trying LRO, so as to avoid skb allocation overheads */
dd50f336
BG
1369
1370 skb = netdev_alloc_skb(dev, MYRI10GE_HLEN + 16);
1371 if (unlikely(skb == NULL)) {
d6279c88 1372 ss->stats.rx_dropped++;
dd50f336
BG
1373 do {
1374 i--;
1375 put_page(rx_frags[i].page);
1376 } while (i != 0);
1377 return 0;
1378 }
1379
1380 /* Attach the pages to the skb, and trim off any padding */
1381 myri10ge_rx_skb_build(skb, va, rx_frags, len, hlen);
1382 if (skb_shinfo(skb)->frags[0].size <= 0) {
1383 put_page(skb_shinfo(skb)->frags[0].page);
1384 skb_shinfo(skb)->nr_frags = 0;
1385 }
1386 skb->protocol = eth_type_trans(skb, dev);
0c8dfc83 1387 skb_record_rx_queue(skb, ss - &mgp->ss[0]);
dd50f336
BG
1388
1389 if (mgp->csum_flag) {
1390 if ((skb->protocol == htons(ETH_P_IP)) ||
1391 (skb->protocol == htons(ETH_P_IPV6))) {
1392 skb->csum = csum;
1393 skb->ip_summed = CHECKSUM_COMPLETE;
1394 } else
1395 myri10ge_vlan_ip_csum(skb, csum);
1396 }
1397 netif_receive_skb(skb);
dd50f336
BG
1398 return 1;
1399}
1400
b53bef84
BG
1401static inline void
1402myri10ge_tx_done(struct myri10ge_slice_state *ss, int mcp_index)
0da34b6d 1403{
b53bef84
BG
1404 struct pci_dev *pdev = ss->mgp->pdev;
1405 struct myri10ge_tx_buf *tx = &ss->tx;
236bb5e6 1406 struct netdev_queue *dev_queue;
0da34b6d
BG
1407 struct sk_buff *skb;
1408 int idx, len;
0da34b6d
BG
1409
1410 while (tx->pkt_done != mcp_index) {
1411 idx = tx->done & tx->mask;
1412 skb = tx->info[idx].skb;
1413
1414 /* Mark as free */
1415 tx->info[idx].skb = NULL;
1416 if (tx->info[idx].last) {
1417 tx->pkt_done++;
1418 tx->info[idx].last = 0;
1419 }
1420 tx->done++;
c755b4b6
FT
1421 len = dma_unmap_len(&tx->info[idx], len);
1422 dma_unmap_len_set(&tx->info[idx], len, 0);
0da34b6d 1423 if (skb) {
b53bef84
BG
1424 ss->stats.tx_bytes += skb->len;
1425 ss->stats.tx_packets++;
0da34b6d
BG
1426 dev_kfree_skb_irq(skb);
1427 if (len)
1428 pci_unmap_single(pdev,
c755b4b6 1429 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
1430 bus), len,
1431 PCI_DMA_TODEVICE);
1432 } else {
1433 if (len)
1434 pci_unmap_page(pdev,
c755b4b6 1435 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
1436 bus), len,
1437 PCI_DMA_TODEVICE);
1438 }
0da34b6d 1439 }
236bb5e6
BG
1440
1441 dev_queue = netdev_get_tx_queue(ss->dev, ss - ss->mgp->ss);
1442 /*
1443 * Make a minimal effort to prevent the NIC from polling an
1444 * idle tx queue. If we can't get the lock we leave the queue
1445 * active. In this case, either a thread was about to start
1446 * using the queue anyway, or we lost a race and the NIC will
1447 * waste some of its resources polling an inactive queue for a
1448 * while.
1449 */
1450
1451 if ((ss->mgp->dev->real_num_tx_queues > 1) &&
1452 __netif_tx_trylock(dev_queue)) {
1453 if (tx->req == tx->done) {
1454 tx->queue_active = 0;
1455 put_be32(htonl(1), tx->send_stop);
8c2f5fa5 1456 mb();
6824a105 1457 mmiowb();
236bb5e6
BG
1458 }
1459 __netif_tx_unlock(dev_queue);
1460 }
1461
0da34b6d 1462 /* start the queue if we've stopped it */
8e95a202
JP
1463 if (netif_tx_queue_stopped(dev_queue) &&
1464 tx->req - tx->done < (tx->mask >> 1)) {
b53bef84 1465 tx->wake_queue++;
236bb5e6 1466 netif_tx_wake_queue(dev_queue);
0da34b6d
BG
1467 }
1468}
1469
b53bef84
BG
1470static inline int
1471myri10ge_clean_rx_done(struct myri10ge_slice_state *ss, int budget)
0da34b6d 1472{
b53bef84
BG
1473 struct myri10ge_rx_done *rx_done = &ss->rx_done;
1474 struct myri10ge_priv *mgp = ss->mgp;
b3cd9657 1475
0da34b6d
BG
1476 unsigned long rx_bytes = 0;
1477 unsigned long rx_packets = 0;
1478 unsigned long rx_ok;
1479
1480 int idx = rx_done->idx;
1481 int cnt = rx_done->cnt;
bea3348e 1482 int work_done = 0;
0da34b6d 1483 u16 length;
40f6cff5 1484 __wsum checksum;
0da34b6d 1485
b3cd9657
SG
1486 /*
1487 * Prevent compiler from generating more than one ->features memory
1488 * access to avoid theoretical race condition with functions that
1489 * change NETIF_F_LRO flag at runtime.
1490 */
1491 bool lro_enabled = ACCESS_ONCE(mgp->dev->features) & NETIF_F_LRO;
1492
c956a240 1493 while (rx_done->entry[idx].length != 0 && work_done < budget) {
0da34b6d
BG
1494 length = ntohs(rx_done->entry[idx].length);
1495 rx_done->entry[idx].length = 0;
40f6cff5 1496 checksum = csum_unfold(rx_done->entry[idx].checksum);
b3cd9657 1497 rx_ok = myri10ge_rx_done(ss, length, checksum, lro_enabled);
0da34b6d
BG
1498 rx_packets += rx_ok;
1499 rx_bytes += rx_ok * (unsigned long)length;
1500 cnt++;
014377a1 1501 idx = cnt & (mgp->max_intr_slots - 1);
c956a240 1502 work_done++;
0da34b6d
BG
1503 }
1504 rx_done->idx = idx;
1505 rx_done->cnt = cnt;
b53bef84
BG
1506 ss->stats.rx_packets += rx_packets;
1507 ss->stats.rx_bytes += rx_bytes;
c7dab99b 1508
b3cd9657 1509 if (lro_enabled)
1e6e9342
AG
1510 lro_flush_all(&rx_done->lro_mgr);
1511
c7dab99b 1512 /* restock receive rings if needed */
b53bef84
BG
1513 if (ss->rx_small.fill_cnt - ss->rx_small.cnt < myri10ge_fill_thresh)
1514 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
c7dab99b 1515 mgp->small_bytes + MXGEFW_PAD, 0);
b53bef84
BG
1516 if (ss->rx_big.fill_cnt - ss->rx_big.cnt < myri10ge_fill_thresh)
1517 myri10ge_alloc_rx_pages(mgp, &ss->rx_big, mgp->big_bytes, 0);
c7dab99b 1518
bea3348e 1519 return work_done;
0da34b6d
BG
1520}
1521
1522static inline void myri10ge_check_statblock(struct myri10ge_priv *mgp)
1523{
0dcffac1 1524 struct mcp_irq_data *stats = mgp->ss[0].fw_stats;
0da34b6d
BG
1525
1526 if (unlikely(stats->stats_updated)) {
798a95db
BG
1527 unsigned link_up = ntohl(stats->link_up);
1528 if (mgp->link_state != link_up) {
1529 mgp->link_state = link_up;
1530
1531 if (mgp->link_state == MXGEFW_LINK_UP) {
c58ac5ca 1532 if (netif_msg_link(mgp))
78ca90ea 1533 netdev_info(mgp->dev, "link up\n");
0da34b6d 1534 netif_carrier_on(mgp->dev);
c58ac5ca 1535 mgp->link_changes++;
0da34b6d 1536 } else {
c58ac5ca 1537 if (netif_msg_link(mgp))
78ca90ea
JP
1538 netdev_info(mgp->dev, "link %s\n",
1539 link_up == MXGEFW_LINK_MYRINET ?
1540 "mismatch (Myrinet detected)" :
1541 "down");
0da34b6d 1542 netif_carrier_off(mgp->dev);
c58ac5ca 1543 mgp->link_changes++;
0da34b6d
BG
1544 }
1545 }
1546 if (mgp->rdma_tags_available !=
b53bef84 1547 ntohl(stats->rdma_tags_available)) {
0da34b6d 1548 mgp->rdma_tags_available =
b53bef84 1549 ntohl(stats->rdma_tags_available);
78ca90ea
JP
1550 netdev_warn(mgp->dev, "RDMA timed out! %d tags left\n",
1551 mgp->rdma_tags_available);
0da34b6d
BG
1552 }
1553 mgp->down_cnt += stats->link_down;
1554 if (stats->link_down)
1555 wake_up(&mgp->down_wq);
1556 }
1557}
1558
bea3348e 1559static int myri10ge_poll(struct napi_struct *napi, int budget)
0da34b6d 1560{
b53bef84
BG
1561 struct myri10ge_slice_state *ss =
1562 container_of(napi, struct myri10ge_slice_state, napi);
bea3348e 1563 int work_done;
0da34b6d 1564
5dd2d332 1565#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
1566 if (ss->mgp->dca_enabled)
1567 myri10ge_update_dca(ss);
1568#endif
1569
0da34b6d 1570 /* process as many rx events as NAPI will allow */
b53bef84 1571 work_done = myri10ge_clean_rx_done(ss, budget);
0da34b6d 1572
4ec24119 1573 if (work_done < budget) {
288379f0 1574 napi_complete(napi);
b53bef84 1575 put_be32(htonl(3), ss->irq_claim);
0da34b6d 1576 }
bea3348e 1577 return work_done;
0da34b6d
BG
1578}
1579
7d12e780 1580static irqreturn_t myri10ge_intr(int irq, void *arg)
0da34b6d 1581{
b53bef84
BG
1582 struct myri10ge_slice_state *ss = arg;
1583 struct myri10ge_priv *mgp = ss->mgp;
1584 struct mcp_irq_data *stats = ss->fw_stats;
1585 struct myri10ge_tx_buf *tx = &ss->tx;
0da34b6d
BG
1586 u32 send_done_count;
1587 int i;
1588
236bb5e6
BG
1589 /* an interrupt on a non-zero receive-only slice is implicitly
1590 * valid since MSI-X irqs are not shared */
1591 if ((mgp->dev->real_num_tx_queues == 1) && (ss != mgp->ss)) {
288379f0 1592 napi_schedule(&ss->napi);
807540ba 1593 return IRQ_HANDLED;
0dcffac1
BG
1594 }
1595
0da34b6d
BG
1596 /* make sure it is our IRQ, and that the DMA has finished */
1597 if (unlikely(!stats->valid))
807540ba 1598 return IRQ_NONE;
0da34b6d
BG
1599
1600 /* low bit indicates receives are present, so schedule
1601 * napi poll handler */
1602 if (stats->valid & 1)
288379f0 1603 napi_schedule(&ss->napi);
0da34b6d 1604
0dcffac1 1605 if (!mgp->msi_enabled && !mgp->msix_enabled) {
40f6cff5 1606 put_be32(0, mgp->irq_deassert);
0da34b6d
BG
1607 if (!myri10ge_deassert_wait)
1608 stats->valid = 0;
1609 mb();
1610 } else
1611 stats->valid = 0;
1612
1613 /* Wait for IRQ line to go low, if using INTx */
1614 i = 0;
1615 while (1) {
1616 i++;
1617 /* check for transmit completes and receives */
1618 send_done_count = ntohl(stats->send_done_count);
1619 if (send_done_count != tx->pkt_done)
b53bef84 1620 myri10ge_tx_done(ss, (int)send_done_count);
0da34b6d 1621 if (unlikely(i > myri10ge_max_irq_loops)) {
78ca90ea 1622 netdev_err(mgp->dev, "irq stuck?\n");
0da34b6d
BG
1623 stats->valid = 0;
1624 schedule_work(&mgp->watchdog_work);
1625 }
1626 if (likely(stats->valid == 0))
1627 break;
1628 cpu_relax();
1629 barrier();
1630 }
1631
236bb5e6
BG
1632 /* Only slice 0 updates stats */
1633 if (ss == mgp->ss)
1634 myri10ge_check_statblock(mgp);
0da34b6d 1635
b53bef84 1636 put_be32(htonl(3), ss->irq_claim + 1);
807540ba 1637 return IRQ_HANDLED;
0da34b6d
BG
1638}
1639
1640static int
1641myri10ge_get_settings(struct net_device *netdev, struct ethtool_cmd *cmd)
1642{
c0bf8801
BG
1643 struct myri10ge_priv *mgp = netdev_priv(netdev);
1644 char *ptr;
1645 int i;
1646
0da34b6d
BG
1647 cmd->autoneg = AUTONEG_DISABLE;
1648 cmd->speed = SPEED_10000;
1649 cmd->duplex = DUPLEX_FULL;
c0bf8801
BG
1650
1651 /*
1652 * parse the product code to deterimine the interface type
1653 * (CX4, XFP, Quad Ribbon Fiber) by looking at the character
1654 * after the 3rd dash in the driver's cached copy of the
1655 * EEPROM's product code string.
1656 */
1657 ptr = mgp->product_code_string;
1658 if (ptr == NULL) {
78ca90ea 1659 netdev_err(netdev, "Missing product code\n");
c0bf8801
BG
1660 return 0;
1661 }
1662 for (i = 0; i < 3; i++, ptr++) {
1663 ptr = strchr(ptr, '-');
1664 if (ptr == NULL) {
78ca90ea
JP
1665 netdev_err(netdev, "Invalid product code %s\n",
1666 mgp->product_code_string);
c0bf8801
BG
1667 return 0;
1668 }
1669 }
196f17eb
BG
1670 if (*ptr == '2')
1671 ptr++;
1672 if (*ptr == 'R' || *ptr == 'Q' || *ptr == 'S') {
1673 /* We've found either an XFP, quad ribbon fiber, or SFP+ */
c0bf8801 1674 cmd->port = PORT_FIBRE;
196f17eb
BG
1675 cmd->supported |= SUPPORTED_FIBRE;
1676 cmd->advertising |= ADVERTISED_FIBRE;
1677 } else {
1678 cmd->port = PORT_OTHER;
c0bf8801 1679 }
196f17eb
BG
1680 if (*ptr == 'R' || *ptr == 'S')
1681 cmd->transceiver = XCVR_EXTERNAL;
1682 else
1683 cmd->transceiver = XCVR_INTERNAL;
1684
0da34b6d
BG
1685 return 0;
1686}
1687
1688static void
1689myri10ge_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *info)
1690{
1691 struct myri10ge_priv *mgp = netdev_priv(netdev);
1692
1693 strlcpy(info->driver, "myri10ge", sizeof(info->driver));
1694 strlcpy(info->version, MYRI10GE_VERSION_STR, sizeof(info->version));
1695 strlcpy(info->fw_version, mgp->fw_version, sizeof(info->fw_version));
1696 strlcpy(info->bus_info, pci_name(mgp->pdev), sizeof(info->bus_info));
1697}
1698
1699static int
1700myri10ge_get_coalesce(struct net_device *netdev, struct ethtool_coalesce *coal)
1701{
1702 struct myri10ge_priv *mgp = netdev_priv(netdev);
99f5f87e 1703
0da34b6d
BG
1704 coal->rx_coalesce_usecs = mgp->intr_coal_delay;
1705 return 0;
1706}
1707
1708static int
1709myri10ge_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *coal)
1710{
1711 struct myri10ge_priv *mgp = netdev_priv(netdev);
1712
1713 mgp->intr_coal_delay = coal->rx_coalesce_usecs;
40f6cff5 1714 put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr);
0da34b6d
BG
1715 return 0;
1716}
1717
1718static void
1719myri10ge_get_pauseparam(struct net_device *netdev,
1720 struct ethtool_pauseparam *pause)
1721{
1722 struct myri10ge_priv *mgp = netdev_priv(netdev);
1723
1724 pause->autoneg = 0;
1725 pause->rx_pause = mgp->pause;
1726 pause->tx_pause = mgp->pause;
1727}
1728
1729static int
1730myri10ge_set_pauseparam(struct net_device *netdev,
1731 struct ethtool_pauseparam *pause)
1732{
1733 struct myri10ge_priv *mgp = netdev_priv(netdev);
1734
1735 if (pause->tx_pause != mgp->pause)
1736 return myri10ge_change_pause(mgp, pause->tx_pause);
1737 if (pause->rx_pause != mgp->pause)
2488f56d 1738 return myri10ge_change_pause(mgp, pause->rx_pause);
0da34b6d
BG
1739 if (pause->autoneg != 0)
1740 return -EINVAL;
1741 return 0;
1742}
1743
1744static void
1745myri10ge_get_ringparam(struct net_device *netdev,
1746 struct ethtool_ringparam *ring)
1747{
1748 struct myri10ge_priv *mgp = netdev_priv(netdev);
1749
0dcffac1
BG
1750 ring->rx_mini_max_pending = mgp->ss[0].rx_small.mask + 1;
1751 ring->rx_max_pending = mgp->ss[0].rx_big.mask + 1;
0da34b6d 1752 ring->rx_jumbo_max_pending = 0;
6498be3f 1753 ring->tx_max_pending = mgp->ss[0].tx.mask + 1;
0da34b6d
BG
1754 ring->rx_mini_pending = ring->rx_mini_max_pending;
1755 ring->rx_pending = ring->rx_max_pending;
1756 ring->rx_jumbo_pending = ring->rx_jumbo_max_pending;
1757 ring->tx_pending = ring->tx_max_pending;
1758}
1759
1760static u32 myri10ge_get_rx_csum(struct net_device *netdev)
1761{
1762 struct myri10ge_priv *mgp = netdev_priv(netdev);
99f5f87e 1763
0da34b6d
BG
1764 if (mgp->csum_flag)
1765 return 1;
1766 else
1767 return 0;
1768}
1769
1770static int myri10ge_set_rx_csum(struct net_device *netdev, u32 csum_enabled)
1771{
1772 struct myri10ge_priv *mgp = netdev_priv(netdev);
3a0c7d2d 1773 int err = 0;
99f5f87e 1774
0da34b6d
BG
1775 if (csum_enabled)
1776 mgp->csum_flag = MXGEFW_FLAGS_CKSUM;
3a0c7d2d 1777 else {
1437ce39 1778 netdev->features &= ~NETIF_F_LRO;
0da34b6d 1779 mgp->csum_flag = 0;
3a0c7d2d
BG
1780
1781 }
1782 return err;
0da34b6d
BG
1783}
1784
4f93fde0
BG
1785static int myri10ge_set_tso(struct net_device *netdev, u32 tso_enabled)
1786{
1787 struct myri10ge_priv *mgp = netdev_priv(netdev);
04ed3e74 1788 u32 flags = mgp->features & (NETIF_F_TSO6 | NETIF_F_TSO);
4f93fde0
BG
1789
1790 if (tso_enabled)
1791 netdev->features |= flags;
1792 else
1793 netdev->features &= ~flags;
1794 return 0;
1795}
1796
b53bef84 1797static const char myri10ge_gstrings_main_stats[][ETH_GSTRING_LEN] = {
0da34b6d
BG
1798 "rx_packets", "tx_packets", "rx_bytes", "tx_bytes", "rx_errors",
1799 "tx_errors", "rx_dropped", "tx_dropped", "multicast", "collisions",
1800 "rx_length_errors", "rx_over_errors", "rx_crc_errors",
1801 "rx_frame_errors", "rx_fifo_errors", "rx_missed_errors",
1802 "tx_aborted_errors", "tx_carrier_errors", "tx_fifo_errors",
1803 "tx_heartbeat_errors", "tx_window_errors",
1804 /* device-specific stats */
0dcffac1 1805 "tx_boundary", "WC", "irq", "MSI", "MSIX",
0da34b6d 1806 "read_dma_bw_MBs", "write_dma_bw_MBs", "read_write_dma_bw_MBs",
b53bef84 1807 "serial_number", "watchdog_resets",
5dd2d332 1808#ifdef CONFIG_MYRI10GE_DCA
9a6b3b54 1809 "dca_capable_firmware", "dca_device_present",
981813d8 1810#endif
c58ac5ca 1811 "link_changes", "link_up", "dropped_link_overflow",
cee505db
BG
1812 "dropped_link_error_or_filtered",
1813 "dropped_pause", "dropped_bad_phy", "dropped_bad_crc32",
1814 "dropped_unicast_filtered", "dropped_multicast_filtered",
0da34b6d 1815 "dropped_runt", "dropped_overrun", "dropped_no_small_buffer",
b53bef84
BG
1816 "dropped_no_big_buffer"
1817};
1818
1819static const char myri10ge_gstrings_slice_stats[][ETH_GSTRING_LEN] = {
1820 "----------- slice ---------",
1821 "tx_pkt_start", "tx_pkt_done", "tx_req", "tx_done",
1822 "rx_small_cnt", "rx_big_cnt",
1823 "wake_queue", "stop_queue", "tx_linearized", "LRO aggregated",
1824 "LRO flushed",
1e6e9342 1825 "LRO avg aggr", "LRO no_desc"
0da34b6d
BG
1826};
1827
1828#define MYRI10GE_NET_STATS_LEN 21
b53bef84
BG
1829#define MYRI10GE_MAIN_STATS_LEN ARRAY_SIZE(myri10ge_gstrings_main_stats)
1830#define MYRI10GE_SLICE_STATS_LEN ARRAY_SIZE(myri10ge_gstrings_slice_stats)
0da34b6d
BG
1831
1832static void
1833myri10ge_get_strings(struct net_device *netdev, u32 stringset, u8 * data)
1834{
0dcffac1
BG
1835 struct myri10ge_priv *mgp = netdev_priv(netdev);
1836 int i;
1837
0da34b6d
BG
1838 switch (stringset) {
1839 case ETH_SS_STATS:
b53bef84
BG
1840 memcpy(data, *myri10ge_gstrings_main_stats,
1841 sizeof(myri10ge_gstrings_main_stats));
1842 data += sizeof(myri10ge_gstrings_main_stats);
0dcffac1
BG
1843 for (i = 0; i < mgp->num_slices; i++) {
1844 memcpy(data, *myri10ge_gstrings_slice_stats,
1845 sizeof(myri10ge_gstrings_slice_stats));
1846 data += sizeof(myri10ge_gstrings_slice_stats);
1847 }
0da34b6d
BG
1848 break;
1849 }
1850}
1851
b9f2c044 1852static int myri10ge_get_sset_count(struct net_device *netdev, int sset)
0da34b6d 1853{
0dcffac1
BG
1854 struct myri10ge_priv *mgp = netdev_priv(netdev);
1855
b9f2c044
JG
1856 switch (sset) {
1857 case ETH_SS_STATS:
0dcffac1
BG
1858 return MYRI10GE_MAIN_STATS_LEN +
1859 mgp->num_slices * MYRI10GE_SLICE_STATS_LEN;
b9f2c044
JG
1860 default:
1861 return -EOPNOTSUPP;
1862 }
0da34b6d
BG
1863}
1864
1865static void
1866myri10ge_get_ethtool_stats(struct net_device *netdev,
1867 struct ethtool_stats *stats, u64 * data)
1868{
1869 struct myri10ge_priv *mgp = netdev_priv(netdev);
b53bef84 1870 struct myri10ge_slice_state *ss;
0dcffac1 1871 int slice;
0da34b6d
BG
1872 int i;
1873
59081825
BG
1874 /* force stats update */
1875 (void)myri10ge_get_stats(netdev);
0da34b6d 1876 for (i = 0; i < MYRI10GE_NET_STATS_LEN; i++)
6dc34941 1877 data[i] = ((unsigned long *)&netdev->stats)[i];
0da34b6d 1878
b53bef84 1879 data[i++] = (unsigned int)mgp->tx_boundary;
276e26c3 1880 data[i++] = (unsigned int)mgp->wc_enabled;
2c1a1088
BG
1881 data[i++] = (unsigned int)mgp->pdev->irq;
1882 data[i++] = (unsigned int)mgp->msi_enabled;
0dcffac1 1883 data[i++] = (unsigned int)mgp->msix_enabled;
0da34b6d
BG
1884 data[i++] = (unsigned int)mgp->read_dma;
1885 data[i++] = (unsigned int)mgp->write_dma;
1886 data[i++] = (unsigned int)mgp->read_write_dma;
1887 data[i++] = (unsigned int)mgp->serial_number;
0da34b6d 1888 data[i++] = (unsigned int)mgp->watchdog_resets;
5dd2d332 1889#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
1890 data[i++] = (unsigned int)(mgp->ss[0].dca_tag != NULL);
1891 data[i++] = (unsigned int)(mgp->dca_enabled);
1892#endif
c58ac5ca 1893 data[i++] = (unsigned int)mgp->link_changes;
b53bef84
BG
1894
1895 /* firmware stats are useful only in the first slice */
0dcffac1 1896 ss = &mgp->ss[0];
b53bef84
BG
1897 data[i++] = (unsigned int)ntohl(ss->fw_stats->link_up);
1898 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_link_overflow);
cee505db 1899 data[i++] =
b53bef84
BG
1900 (unsigned int)ntohl(ss->fw_stats->dropped_link_error_or_filtered);
1901 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_pause);
1902 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_bad_phy);
1903 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_bad_crc32);
1904 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_unicast_filtered);
85a7ea1b 1905 data[i++] =
b53bef84
BG
1906 (unsigned int)ntohl(ss->fw_stats->dropped_multicast_filtered);
1907 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_runt);
1908 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_overrun);
1909 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_no_small_buffer);
1910 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_no_big_buffer);
1911
0dcffac1
BG
1912 for (slice = 0; slice < mgp->num_slices; slice++) {
1913 ss = &mgp->ss[slice];
1914 data[i++] = slice;
1915 data[i++] = (unsigned int)ss->tx.pkt_start;
1916 data[i++] = (unsigned int)ss->tx.pkt_done;
1917 data[i++] = (unsigned int)ss->tx.req;
1918 data[i++] = (unsigned int)ss->tx.done;
1919 data[i++] = (unsigned int)ss->rx_small.cnt;
1920 data[i++] = (unsigned int)ss->rx_big.cnt;
1921 data[i++] = (unsigned int)ss->tx.wake_queue;
1922 data[i++] = (unsigned int)ss->tx.stop_queue;
1923 data[i++] = (unsigned int)ss->tx.linearized;
1924 data[i++] = ss->rx_done.lro_mgr.stats.aggregated;
1925 data[i++] = ss->rx_done.lro_mgr.stats.flushed;
1926 if (ss->rx_done.lro_mgr.stats.flushed)
1927 data[i++] = ss->rx_done.lro_mgr.stats.aggregated /
1928 ss->rx_done.lro_mgr.stats.flushed;
1929 else
1930 data[i++] = 0;
1931 data[i++] = ss->rx_done.lro_mgr.stats.no_desc;
1932 }
0da34b6d
BG
1933}
1934
c58ac5ca
BG
1935static void myri10ge_set_msglevel(struct net_device *netdev, u32 value)
1936{
1937 struct myri10ge_priv *mgp = netdev_priv(netdev);
1938 mgp->msg_enable = value;
1939}
1940
1941static u32 myri10ge_get_msglevel(struct net_device *netdev)
1942{
1943 struct myri10ge_priv *mgp = netdev_priv(netdev);
1944 return mgp->msg_enable;
1945}
1946
1437ce39
BH
1947static int myri10ge_set_flags(struct net_device *netdev, u32 value)
1948{
1949 return ethtool_op_set_flags(netdev, value, ETH_FLAG_LRO);
1950}
1951
7282d491 1952static const struct ethtool_ops myri10ge_ethtool_ops = {
0da34b6d
BG
1953 .get_settings = myri10ge_get_settings,
1954 .get_drvinfo = myri10ge_get_drvinfo,
1955 .get_coalesce = myri10ge_get_coalesce,
1956 .set_coalesce = myri10ge_set_coalesce,
1957 .get_pauseparam = myri10ge_get_pauseparam,
1958 .set_pauseparam = myri10ge_set_pauseparam,
1959 .get_ringparam = myri10ge_get_ringparam,
1960 .get_rx_csum = myri10ge_get_rx_csum,
1961 .set_rx_csum = myri10ge_set_rx_csum,
b10c0668 1962 .set_tx_csum = ethtool_op_set_tx_hw_csum,
0da34b6d 1963 .set_sg = ethtool_op_set_sg,
4f93fde0 1964 .set_tso = myri10ge_set_tso,
6ffdd071 1965 .get_link = ethtool_op_get_link,
0da34b6d 1966 .get_strings = myri10ge_get_strings,
b9f2c044 1967 .get_sset_count = myri10ge_get_sset_count,
c58ac5ca
BG
1968 .get_ethtool_stats = myri10ge_get_ethtool_stats,
1969 .set_msglevel = myri10ge_set_msglevel,
3a0c7d2d
BG
1970 .get_msglevel = myri10ge_get_msglevel,
1971 .get_flags = ethtool_op_get_flags,
1437ce39 1972 .set_flags = myri10ge_set_flags
0da34b6d
BG
1973};
1974
b53bef84 1975static int myri10ge_allocate_rings(struct myri10ge_slice_state *ss)
0da34b6d 1976{
b53bef84 1977 struct myri10ge_priv *mgp = ss->mgp;
0da34b6d 1978 struct myri10ge_cmd cmd;
b53bef84 1979 struct net_device *dev = mgp->dev;
0da34b6d
BG
1980 int tx_ring_size, rx_ring_size;
1981 int tx_ring_entries, rx_ring_entries;
0dcffac1 1982 int i, slice, status;
0da34b6d
BG
1983 size_t bytes;
1984
0da34b6d 1985 /* get ring sizes */
0dcffac1
BG
1986 slice = ss - mgp->ss;
1987 cmd.data0 = slice;
0da34b6d
BG
1988 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_RING_SIZE, &cmd, 0);
1989 tx_ring_size = cmd.data0;
0dcffac1 1990 cmd.data0 = slice;
0da34b6d 1991 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_RX_RING_SIZE, &cmd, 0);
355c7265
BG
1992 if (status != 0)
1993 return status;
0da34b6d
BG
1994 rx_ring_size = cmd.data0;
1995
1996 tx_ring_entries = tx_ring_size / sizeof(struct mcp_kreq_ether_send);
1997 rx_ring_entries = rx_ring_size / sizeof(struct mcp_dma_addr);
b53bef84
BG
1998 ss->tx.mask = tx_ring_entries - 1;
1999 ss->rx_small.mask = ss->rx_big.mask = rx_ring_entries - 1;
0da34b6d 2000
355c7265
BG
2001 status = -ENOMEM;
2002
0da34b6d
BG
2003 /* allocate the host shadow rings */
2004
2005 bytes = 8 + (MYRI10GE_MAX_SEND_DESC_TSO + 4)
b53bef84
BG
2006 * sizeof(*ss->tx.req_list);
2007 ss->tx.req_bytes = kzalloc(bytes, GFP_KERNEL);
2008 if (ss->tx.req_bytes == NULL)
0da34b6d
BG
2009 goto abort_with_nothing;
2010
2011 /* ensure req_list entries are aligned to 8 bytes */
b53bef84
BG
2012 ss->tx.req_list = (struct mcp_kreq_ether_send *)
2013 ALIGN((unsigned long)ss->tx.req_bytes, 8);
236bb5e6 2014 ss->tx.queue_active = 0;
0da34b6d 2015
b53bef84
BG
2016 bytes = rx_ring_entries * sizeof(*ss->rx_small.shadow);
2017 ss->rx_small.shadow = kzalloc(bytes, GFP_KERNEL);
2018 if (ss->rx_small.shadow == NULL)
0da34b6d
BG
2019 goto abort_with_tx_req_bytes;
2020
b53bef84
BG
2021 bytes = rx_ring_entries * sizeof(*ss->rx_big.shadow);
2022 ss->rx_big.shadow = kzalloc(bytes, GFP_KERNEL);
2023 if (ss->rx_big.shadow == NULL)
0da34b6d
BG
2024 goto abort_with_rx_small_shadow;
2025
2026 /* allocate the host info rings */
2027
b53bef84
BG
2028 bytes = tx_ring_entries * sizeof(*ss->tx.info);
2029 ss->tx.info = kzalloc(bytes, GFP_KERNEL);
2030 if (ss->tx.info == NULL)
0da34b6d
BG
2031 goto abort_with_rx_big_shadow;
2032
b53bef84
BG
2033 bytes = rx_ring_entries * sizeof(*ss->rx_small.info);
2034 ss->rx_small.info = kzalloc(bytes, GFP_KERNEL);
2035 if (ss->rx_small.info == NULL)
0da34b6d
BG
2036 goto abort_with_tx_info;
2037
b53bef84
BG
2038 bytes = rx_ring_entries * sizeof(*ss->rx_big.info);
2039 ss->rx_big.info = kzalloc(bytes, GFP_KERNEL);
2040 if (ss->rx_big.info == NULL)
0da34b6d
BG
2041 goto abort_with_rx_small_info;
2042
2043 /* Fill the receive rings */
b53bef84
BG
2044 ss->rx_big.cnt = 0;
2045 ss->rx_small.cnt = 0;
2046 ss->rx_big.fill_cnt = 0;
2047 ss->rx_small.fill_cnt = 0;
2048 ss->rx_small.page_offset = MYRI10GE_ALLOC_SIZE;
2049 ss->rx_big.page_offset = MYRI10GE_ALLOC_SIZE;
2050 ss->rx_small.watchdog_needed = 0;
2051 ss->rx_big.watchdog_needed = 0;
2052 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
c7dab99b 2053 mgp->small_bytes + MXGEFW_PAD, 0);
0da34b6d 2054
b53bef84 2055 if (ss->rx_small.fill_cnt < ss->rx_small.mask + 1) {
78ca90ea
JP
2056 netdev_err(dev, "slice-%d: alloced only %d small bufs\n",
2057 slice, ss->rx_small.fill_cnt);
c7dab99b 2058 goto abort_with_rx_small_ring;
0da34b6d
BG
2059 }
2060
b53bef84
BG
2061 myri10ge_alloc_rx_pages(mgp, &ss->rx_big, mgp->big_bytes, 0);
2062 if (ss->rx_big.fill_cnt < ss->rx_big.mask + 1) {
78ca90ea
JP
2063 netdev_err(dev, "slice-%d: alloced only %d big bufs\n",
2064 slice, ss->rx_big.fill_cnt);
c7dab99b 2065 goto abort_with_rx_big_ring;
0da34b6d
BG
2066 }
2067
2068 return 0;
2069
2070abort_with_rx_big_ring:
b53bef84
BG
2071 for (i = ss->rx_big.cnt; i < ss->rx_big.fill_cnt; i++) {
2072 int idx = i & ss->rx_big.mask;
2073 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_big.info[idx],
c7dab99b 2074 mgp->big_bytes);
b53bef84 2075 put_page(ss->rx_big.info[idx].page);
0da34b6d
BG
2076 }
2077
2078abort_with_rx_small_ring:
b53bef84
BG
2079 for (i = ss->rx_small.cnt; i < ss->rx_small.fill_cnt; i++) {
2080 int idx = i & ss->rx_small.mask;
2081 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_small.info[idx],
c7dab99b 2082 mgp->small_bytes + MXGEFW_PAD);
b53bef84 2083 put_page(ss->rx_small.info[idx].page);
0da34b6d 2084 }
c7dab99b 2085
b53bef84 2086 kfree(ss->rx_big.info);
0da34b6d
BG
2087
2088abort_with_rx_small_info:
b53bef84 2089 kfree(ss->rx_small.info);
0da34b6d
BG
2090
2091abort_with_tx_info:
b53bef84 2092 kfree(ss->tx.info);
0da34b6d
BG
2093
2094abort_with_rx_big_shadow:
b53bef84 2095 kfree(ss->rx_big.shadow);
0da34b6d
BG
2096
2097abort_with_rx_small_shadow:
b53bef84 2098 kfree(ss->rx_small.shadow);
0da34b6d
BG
2099
2100abort_with_tx_req_bytes:
b53bef84
BG
2101 kfree(ss->tx.req_bytes);
2102 ss->tx.req_bytes = NULL;
2103 ss->tx.req_list = NULL;
0da34b6d
BG
2104
2105abort_with_nothing:
2106 return status;
2107}
2108
b53bef84 2109static void myri10ge_free_rings(struct myri10ge_slice_state *ss)
0da34b6d 2110{
b53bef84 2111 struct myri10ge_priv *mgp = ss->mgp;
0da34b6d
BG
2112 struct sk_buff *skb;
2113 struct myri10ge_tx_buf *tx;
2114 int i, len, idx;
2115
0dcffac1
BG
2116 /* If not allocated, skip it */
2117 if (ss->tx.req_list == NULL)
2118 return;
2119
b53bef84
BG
2120 for (i = ss->rx_big.cnt; i < ss->rx_big.fill_cnt; i++) {
2121 idx = i & ss->rx_big.mask;
2122 if (i == ss->rx_big.fill_cnt - 1)
2123 ss->rx_big.info[idx].page_offset = MYRI10GE_ALLOC_SIZE;
2124 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_big.info[idx],
c7dab99b 2125 mgp->big_bytes);
b53bef84 2126 put_page(ss->rx_big.info[idx].page);
0da34b6d
BG
2127 }
2128
b53bef84
BG
2129 for (i = ss->rx_small.cnt; i < ss->rx_small.fill_cnt; i++) {
2130 idx = i & ss->rx_small.mask;
2131 if (i == ss->rx_small.fill_cnt - 1)
2132 ss->rx_small.info[idx].page_offset =
c7dab99b 2133 MYRI10GE_ALLOC_SIZE;
b53bef84 2134 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_small.info[idx],
c7dab99b 2135 mgp->small_bytes + MXGEFW_PAD);
b53bef84 2136 put_page(ss->rx_small.info[idx].page);
c7dab99b 2137 }
b53bef84 2138 tx = &ss->tx;
0da34b6d
BG
2139 while (tx->done != tx->req) {
2140 idx = tx->done & tx->mask;
2141 skb = tx->info[idx].skb;
2142
2143 /* Mark as free */
2144 tx->info[idx].skb = NULL;
2145 tx->done++;
c755b4b6
FT
2146 len = dma_unmap_len(&tx->info[idx], len);
2147 dma_unmap_len_set(&tx->info[idx], len, 0);
0da34b6d 2148 if (skb) {
b53bef84 2149 ss->stats.tx_dropped++;
0da34b6d
BG
2150 dev_kfree_skb_any(skb);
2151 if (len)
2152 pci_unmap_single(mgp->pdev,
c755b4b6 2153 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
2154 bus), len,
2155 PCI_DMA_TODEVICE);
2156 } else {
2157 if (len)
2158 pci_unmap_page(mgp->pdev,
c755b4b6 2159 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
2160 bus), len,
2161 PCI_DMA_TODEVICE);
2162 }
2163 }
b53bef84 2164 kfree(ss->rx_big.info);
0da34b6d 2165
b53bef84 2166 kfree(ss->rx_small.info);
0da34b6d 2167
b53bef84 2168 kfree(ss->tx.info);
0da34b6d 2169
b53bef84 2170 kfree(ss->rx_big.shadow);
0da34b6d 2171
b53bef84 2172 kfree(ss->rx_small.shadow);
0da34b6d 2173
b53bef84
BG
2174 kfree(ss->tx.req_bytes);
2175 ss->tx.req_bytes = NULL;
2176 ss->tx.req_list = NULL;
0da34b6d
BG
2177}
2178
df30a740
BG
2179static int myri10ge_request_irq(struct myri10ge_priv *mgp)
2180{
2181 struct pci_dev *pdev = mgp->pdev;
0dcffac1
BG
2182 struct myri10ge_slice_state *ss;
2183 struct net_device *netdev = mgp->dev;
2184 int i;
df30a740
BG
2185 int status;
2186
0dcffac1
BG
2187 mgp->msi_enabled = 0;
2188 mgp->msix_enabled = 0;
2189 status = 0;
df30a740 2190 if (myri10ge_msi) {
0dcffac1
BG
2191 if (mgp->num_slices > 1) {
2192 status =
2193 pci_enable_msix(pdev, mgp->msix_vectors,
2194 mgp->num_slices);
2195 if (status == 0) {
2196 mgp->msix_enabled = 1;
2197 } else {
2198 dev_err(&pdev->dev,
2199 "Error %d setting up MSI-X\n", status);
2200 return status;
2201 }
2202 }
2203 if (mgp->msix_enabled == 0) {
2204 status = pci_enable_msi(pdev);
2205 if (status != 0) {
2206 dev_err(&pdev->dev,
2207 "Error %d setting up MSI; falling back to xPIC\n",
2208 status);
2209 } else {
2210 mgp->msi_enabled = 1;
2211 }
2212 }
df30a740 2213 }
0dcffac1
BG
2214 if (mgp->msix_enabled) {
2215 for (i = 0; i < mgp->num_slices; i++) {
2216 ss = &mgp->ss[i];
2217 snprintf(ss->irq_desc, sizeof(ss->irq_desc),
2218 "%s:slice-%d", netdev->name, i);
2219 status = request_irq(mgp->msix_vectors[i].vector,
2220 myri10ge_intr, 0, ss->irq_desc,
2221 ss);
2222 if (status != 0) {
2223 dev_err(&pdev->dev,
2224 "slice %d failed to allocate IRQ\n", i);
2225 i--;
2226 while (i >= 0) {
2227 free_irq(mgp->msix_vectors[i].vector,
2228 &mgp->ss[i]);
2229 i--;
2230 }
2231 pci_disable_msix(pdev);
2232 return status;
2233 }
2234 }
2235 } else {
2236 status = request_irq(pdev->irq, myri10ge_intr, IRQF_SHARED,
2237 mgp->dev->name, &mgp->ss[0]);
2238 if (status != 0) {
2239 dev_err(&pdev->dev, "failed to allocate IRQ\n");
2240 if (mgp->msi_enabled)
2241 pci_disable_msi(pdev);
2242 }
df30a740
BG
2243 }
2244 return status;
2245}
2246
2247static void myri10ge_free_irq(struct myri10ge_priv *mgp)
2248{
2249 struct pci_dev *pdev = mgp->pdev;
0dcffac1 2250 int i;
df30a740 2251
0dcffac1
BG
2252 if (mgp->msix_enabled) {
2253 for (i = 0; i < mgp->num_slices; i++)
2254 free_irq(mgp->msix_vectors[i].vector, &mgp->ss[i]);
2255 } else {
2256 free_irq(pdev->irq, &mgp->ss[0]);
2257 }
df30a740
BG
2258 if (mgp->msi_enabled)
2259 pci_disable_msi(pdev);
0dcffac1
BG
2260 if (mgp->msix_enabled)
2261 pci_disable_msix(pdev);
df30a740
BG
2262}
2263
1e6e9342
AG
2264static int
2265myri10ge_get_frag_header(struct skb_frag_struct *frag, void **mac_hdr,
2266 void **ip_hdr, void **tcpudp_hdr,
2267 u64 * hdr_flags, void *priv)
2268{
2269 struct ethhdr *eh;
2270 struct vlan_ethhdr *veh;
2271 struct iphdr *iph;
2272 u8 *va = page_address(frag->page) + frag->page_offset;
2273 unsigned long ll_hlen;
66341fff
AV
2274 /* passed opaque through lro_receive_frags() */
2275 __wsum csum = (__force __wsum) (unsigned long)priv;
1e6e9342
AG
2276
2277 /* find the mac header, aborting if not IPv4 */
2278
2279 eh = (struct ethhdr *)va;
2280 *mac_hdr = eh;
2281 ll_hlen = ETH_HLEN;
2282 if (eh->h_proto != htons(ETH_P_IP)) {
2283 if (eh->h_proto == htons(ETH_P_8021Q)) {
2284 veh = (struct vlan_ethhdr *)va;
2285 if (veh->h_vlan_encapsulated_proto != htons(ETH_P_IP))
2286 return -1;
2287
2288 ll_hlen += VLAN_HLEN;
2289
2290 /*
2291 * HW checksum starts ETH_HLEN bytes into
2292 * frame, so we must subtract off the VLAN
2293 * header's checksum before csum can be used
2294 */
2295 csum = csum_sub(csum, csum_partial(va + ETH_HLEN,
2296 VLAN_HLEN, 0));
2297 } else {
2298 return -1;
2299 }
2300 }
2301 *hdr_flags = LRO_IPV4;
2302
2303 iph = (struct iphdr *)(va + ll_hlen);
2304 *ip_hdr = iph;
2305 if (iph->protocol != IPPROTO_TCP)
2306 return -1;
bcb09dc2
BG
2307 if (iph->frag_off & htons(IP_MF | IP_OFFSET))
2308 return -1;
1e6e9342
AG
2309 *hdr_flags |= LRO_TCP;
2310 *tcpudp_hdr = (u8 *) (*ip_hdr) + (iph->ihl << 2);
2311
2312 /* verify the IP checksum */
2313 if (unlikely(ip_fast_csum((u8 *) iph, iph->ihl)))
2314 return -1;
2315
2316 /* verify the checksum */
2317 if (unlikely(csum_tcpudp_magic(iph->saddr, iph->daddr,
2318 ntohs(iph->tot_len) - (iph->ihl << 2),
2319 IPPROTO_TCP, csum)))
2320 return -1;
2321
2322 return 0;
2323}
2324
77929732
BG
2325static int myri10ge_get_txrx(struct myri10ge_priv *mgp, int slice)
2326{
2327 struct myri10ge_cmd cmd;
2328 struct myri10ge_slice_state *ss;
2329 int status;
2330
2331 ss = &mgp->ss[slice];
236bb5e6
BG
2332 status = 0;
2333 if (slice == 0 || (mgp->dev->real_num_tx_queues > 1)) {
2334 cmd.data0 = slice;
2335 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_OFFSET,
2336 &cmd, 0);
2337 ss->tx.lanai = (struct mcp_kreq_ether_send __iomem *)
2338 (mgp->sram + cmd.data0);
2339 }
77929732
BG
2340 cmd.data0 = slice;
2341 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SMALL_RX_OFFSET,
2342 &cmd, 0);
2343 ss->rx_small.lanai = (struct mcp_kreq_ether_recv __iomem *)
2344 (mgp->sram + cmd.data0);
2345
2346 cmd.data0 = slice;
2347 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_BIG_RX_OFFSET, &cmd, 0);
2348 ss->rx_big.lanai = (struct mcp_kreq_ether_recv __iomem *)
2349 (mgp->sram + cmd.data0);
2350
236bb5e6
BG
2351 ss->tx.send_go = (__iomem __be32 *)
2352 (mgp->sram + MXGEFW_ETH_SEND_GO + 64 * slice);
2353 ss->tx.send_stop = (__iomem __be32 *)
2354 (mgp->sram + MXGEFW_ETH_SEND_STOP + 64 * slice);
77929732
BG
2355 return status;
2356
2357}
2358
2359static int myri10ge_set_stats(struct myri10ge_priv *mgp, int slice)
2360{
2361 struct myri10ge_cmd cmd;
2362 struct myri10ge_slice_state *ss;
2363 int status;
2364
2365 ss = &mgp->ss[slice];
2366 cmd.data0 = MYRI10GE_LOWPART_TO_U32(ss->fw_stats_bus);
2367 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(ss->fw_stats_bus);
236bb5e6 2368 cmd.data2 = sizeof(struct mcp_irq_data) | (slice << 16);
77929732
BG
2369 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_STATS_DMA_V2, &cmd, 0);
2370 if (status == -ENOSYS) {
2371 dma_addr_t bus = ss->fw_stats_bus;
2372 if (slice != 0)
2373 return -EINVAL;
2374 bus += offsetof(struct mcp_irq_data, send_done_count);
2375 cmd.data0 = MYRI10GE_LOWPART_TO_U32(bus);
2376 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(bus);
2377 status = myri10ge_send_cmd(mgp,
2378 MXGEFW_CMD_SET_STATS_DMA_OBSOLETE,
2379 &cmd, 0);
2380 /* Firmware cannot support multicast without STATS_DMA_V2 */
2381 mgp->fw_multicast_support = 0;
2382 } else {
2383 mgp->fw_multicast_support = 1;
2384 }
2385 return 0;
2386}
77929732 2387
0da34b6d
BG
2388static int myri10ge_open(struct net_device *dev)
2389{
0dcffac1 2390 struct myri10ge_slice_state *ss;
b53bef84 2391 struct myri10ge_priv *mgp = netdev_priv(dev);
0da34b6d 2392 struct myri10ge_cmd cmd;
0dcffac1
BG
2393 int i, status, big_pow2, slice;
2394 u8 *itable;
1e6e9342 2395 struct net_lro_mgr *lro_mgr;
0da34b6d 2396
0da34b6d
BG
2397 if (mgp->running != MYRI10GE_ETH_STOPPED)
2398 return -EBUSY;
2399
2400 mgp->running = MYRI10GE_ETH_STARTING;
2401 status = myri10ge_reset(mgp);
2402 if (status != 0) {
78ca90ea 2403 netdev_err(dev, "failed reset\n");
df30a740 2404 goto abort_with_nothing;
0da34b6d
BG
2405 }
2406
0dcffac1
BG
2407 if (mgp->num_slices > 1) {
2408 cmd.data0 = mgp->num_slices;
236bb5e6
BG
2409 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
2410 if (mgp->dev->real_num_tx_queues > 1)
2411 cmd.data1 |= MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES;
0dcffac1
BG
2412 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ENABLE_RSS_QUEUES,
2413 &cmd, 0);
2414 if (status != 0) {
78ca90ea 2415 netdev_err(dev, "failed to set number of slices\n");
0dcffac1
BG
2416 goto abort_with_nothing;
2417 }
2418 /* setup the indirection table */
2419 cmd.data0 = mgp->num_slices;
2420 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_TABLE_SIZE,
2421 &cmd, 0);
2422
2423 status |= myri10ge_send_cmd(mgp,
2424 MXGEFW_CMD_GET_RSS_TABLE_OFFSET,
2425 &cmd, 0);
2426 if (status != 0) {
78ca90ea 2427 netdev_err(dev, "failed to setup rss tables\n");
236bb5e6 2428 goto abort_with_nothing;
0dcffac1
BG
2429 }
2430
2431 /* just enable an identity mapping */
2432 itable = mgp->sram + cmd.data0;
2433 for (i = 0; i < mgp->num_slices; i++)
2434 __raw_writeb(i, &itable[i]);
2435
2436 cmd.data0 = 1;
2437 cmd.data1 = myri10ge_rss_hash;
2438 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_ENABLE,
2439 &cmd, 0);
2440 if (status != 0) {
78ca90ea 2441 netdev_err(dev, "failed to enable slices\n");
0dcffac1
BG
2442 goto abort_with_nothing;
2443 }
2444 }
2445
df30a740
BG
2446 status = myri10ge_request_irq(mgp);
2447 if (status != 0)
2448 goto abort_with_nothing;
2449
0da34b6d
BG
2450 /* decide what small buffer size to use. For good TCP rx
2451 * performance, it is important to not receive 1514 byte
2452 * frames into jumbo buffers, as it confuses the socket buffer
2453 * accounting code, leading to drops and erratic performance.
2454 */
2455
2456 if (dev->mtu <= ETH_DATA_LEN)
c7dab99b
BG
2457 /* enough for a TCP header */
2458 mgp->small_bytes = (128 > SMP_CACHE_BYTES)
2459 ? (128 - MXGEFW_PAD)
2460 : (SMP_CACHE_BYTES - MXGEFW_PAD);
0da34b6d 2461 else
de3c4507
BG
2462 /* enough for a vlan encapsulated ETH_DATA_LEN frame */
2463 mgp->small_bytes = VLAN_ETH_FRAME_LEN;
0da34b6d
BG
2464
2465 /* Override the small buffer size? */
2466 if (myri10ge_small_bytes > 0)
2467 mgp->small_bytes = myri10ge_small_bytes;
2468
0da34b6d
BG
2469 /* Firmware needs the big buff size as a power of 2. Lie and
2470 * tell him the buffer is larger, because we only use 1
2471 * buffer/pkt, and the mtu will prevent overruns.
2472 */
13348bee 2473 big_pow2 = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
c7dab99b 2474 if (big_pow2 < MYRI10GE_ALLOC_SIZE / 2) {
199126a2 2475 while (!is_power_of_2(big_pow2))
c7dab99b 2476 big_pow2++;
13348bee 2477 mgp->big_bytes = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
c7dab99b
BG
2478 } else {
2479 big_pow2 = MYRI10GE_ALLOC_SIZE;
2480 mgp->big_bytes = big_pow2;
2481 }
2482
0dcffac1
BG
2483 /* setup the per-slice data structures */
2484 for (slice = 0; slice < mgp->num_slices; slice++) {
2485 ss = &mgp->ss[slice];
2486
2487 status = myri10ge_get_txrx(mgp, slice);
2488 if (status != 0) {
78ca90ea 2489 netdev_err(dev, "failed to get ring sizes or locations\n");
0dcffac1
BG
2490 goto abort_with_rings;
2491 }
2492 status = myri10ge_allocate_rings(ss);
2493 if (status != 0)
2494 goto abort_with_rings;
236bb5e6
BG
2495
2496 /* only firmware which supports multiple TX queues
2497 * supports setting up the tx stats on non-zero
2498 * slices */
2499 if (slice == 0 || mgp->dev->real_num_tx_queues > 1)
0dcffac1
BG
2500 status = myri10ge_set_stats(mgp, slice);
2501 if (status) {
78ca90ea 2502 netdev_err(dev, "Couldn't set stats DMA\n");
0dcffac1
BG
2503 goto abort_with_rings;
2504 }
2505
2506 lro_mgr = &ss->rx_done.lro_mgr;
2507 lro_mgr->dev = dev;
2508 lro_mgr->features = LRO_F_NAPI;
2509 lro_mgr->ip_summed = CHECKSUM_COMPLETE;
2510 lro_mgr->ip_summed_aggr = CHECKSUM_UNNECESSARY;
2511 lro_mgr->max_desc = MYRI10GE_MAX_LRO_DESCRIPTORS;
2512 lro_mgr->lro_arr = ss->rx_done.lro_desc;
2513 lro_mgr->get_frag_header = myri10ge_get_frag_header;
2514 lro_mgr->max_aggr = myri10ge_lro_max_pkts;
636d2f68 2515 lro_mgr->frag_align_pad = 2;
0dcffac1
BG
2516 if (lro_mgr->max_aggr > MAX_SKB_FRAGS)
2517 lro_mgr->max_aggr = MAX_SKB_FRAGS;
2518
2519 /* must happen prior to any irq */
2520 napi_enable(&(ss)->napi);
2521 }
0da34b6d
BG
2522
2523 /* now give firmware buffers sizes, and MTU */
2524 cmd.data0 = dev->mtu + ETH_HLEN + VLAN_HLEN;
2525 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_MTU, &cmd, 0);
2526 cmd.data0 = mgp->small_bytes;
2527 status |=
2528 myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_SMALL_BUFFER_SIZE, &cmd, 0);
2529 cmd.data0 = big_pow2;
2530 status |=
2531 myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_BIG_BUFFER_SIZE, &cmd, 0);
2532 if (status) {
78ca90ea 2533 netdev_err(dev, "Couldn't set buffer sizes\n");
0da34b6d
BG
2534 goto abort_with_rings;
2535 }
2536
0dcffac1
BG
2537 /*
2538 * Set Linux style TSO mode; this is needed only on newer
2539 * firmware versions. Older versions default to Linux
2540 * style TSO
2541 */
2542 cmd.data0 = 0;
2543 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_TSO_MODE, &cmd, 0);
2544 if (status && status != -ENOSYS) {
78ca90ea 2545 netdev_err(dev, "Couldn't set TSO mode\n");
0da34b6d
BG
2546 goto abort_with_rings;
2547 }
2548
66341fff 2549 mgp->link_state = ~0U;
0da34b6d
BG
2550 mgp->rdma_tags_available = 15;
2551
0da34b6d
BG
2552 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ETHERNET_UP, &cmd, 0);
2553 if (status) {
78ca90ea 2554 netdev_err(dev, "Couldn't bring up link\n");
0da34b6d
BG
2555 goto abort_with_rings;
2556 }
2557
0da34b6d
BG
2558 mgp->running = MYRI10GE_ETH_RUNNING;
2559 mgp->watchdog_timer.expires = jiffies + myri10ge_watchdog_timeout * HZ;
2560 add_timer(&mgp->watchdog_timer);
236bb5e6
BG
2561 netif_tx_wake_all_queues(dev);
2562
0da34b6d
BG
2563 return 0;
2564
2565abort_with_rings:
051d36f3
BG
2566 while (slice) {
2567 slice--;
2568 napi_disable(&mgp->ss[slice].napi);
2569 }
0dcffac1
BG
2570 for (i = 0; i < mgp->num_slices; i++)
2571 myri10ge_free_rings(&mgp->ss[i]);
0da34b6d 2572
df30a740
BG
2573 myri10ge_free_irq(mgp);
2574
0da34b6d
BG
2575abort_with_nothing:
2576 mgp->running = MYRI10GE_ETH_STOPPED;
2577 return -ENOMEM;
2578}
2579
2580static int myri10ge_close(struct net_device *dev)
2581{
b53bef84 2582 struct myri10ge_priv *mgp = netdev_priv(dev);
0da34b6d
BG
2583 struct myri10ge_cmd cmd;
2584 int status, old_down_cnt;
0dcffac1 2585 int i;
0da34b6d 2586
0da34b6d
BG
2587 if (mgp->running != MYRI10GE_ETH_RUNNING)
2588 return 0;
2589
0dcffac1 2590 if (mgp->ss[0].tx.req_bytes == NULL)
0da34b6d
BG
2591 return 0;
2592
2593 del_timer_sync(&mgp->watchdog_timer);
2594 mgp->running = MYRI10GE_ETH_STOPPING;
0dcffac1
BG
2595 for (i = 0; i < mgp->num_slices; i++) {
2596 napi_disable(&mgp->ss[i].napi);
2597 }
0da34b6d 2598 netif_carrier_off(dev);
236bb5e6
BG
2599
2600 netif_tx_stop_all_queues(dev);
d0234215
BG
2601 if (mgp->rebooted == 0) {
2602 old_down_cnt = mgp->down_cnt;
2603 mb();
2604 status =
2605 myri10ge_send_cmd(mgp, MXGEFW_CMD_ETHERNET_DOWN, &cmd, 0);
2606 if (status)
78ca90ea 2607 netdev_err(dev, "Couldn't bring down link\n");
0da34b6d 2608
d0234215
BG
2609 wait_event_timeout(mgp->down_wq, old_down_cnt != mgp->down_cnt,
2610 HZ);
2611 if (old_down_cnt == mgp->down_cnt)
78ca90ea 2612 netdev_err(dev, "never got down irq\n");
d0234215 2613 }
0da34b6d 2614 netif_tx_disable(dev);
df30a740 2615 myri10ge_free_irq(mgp);
0dcffac1
BG
2616 for (i = 0; i < mgp->num_slices; i++)
2617 myri10ge_free_rings(&mgp->ss[i]);
0da34b6d
BG
2618
2619 mgp->running = MYRI10GE_ETH_STOPPED;
2620 return 0;
2621}
2622
2623/* copy an array of struct mcp_kreq_ether_send's to the mcp. Copy
2624 * backwards one at a time and handle ring wraps */
2625
2626static inline void
2627myri10ge_submit_req_backwards(struct myri10ge_tx_buf *tx,
2628 struct mcp_kreq_ether_send *src, int cnt)
2629{
2630 int idx, starting_slot;
2631 starting_slot = tx->req;
2632 while (cnt > 1) {
2633 cnt--;
2634 idx = (starting_slot + cnt) & tx->mask;
2635 myri10ge_pio_copy(&tx->lanai[idx], &src[cnt], sizeof(*src));
2636 mb();
2637 }
2638}
2639
2640/*
2641 * copy an array of struct mcp_kreq_ether_send's to the mcp. Copy
2642 * at most 32 bytes at a time, so as to avoid involving the software
2643 * pio handler in the nic. We re-write the first segment's flags
2644 * to mark them valid only after writing the entire chain.
2645 */
2646
2647static inline void
2648myri10ge_submit_req(struct myri10ge_tx_buf *tx, struct mcp_kreq_ether_send *src,
2649 int cnt)
2650{
2651 int idx, i;
2652 struct mcp_kreq_ether_send __iomem *dstp, *dst;
2653 struct mcp_kreq_ether_send *srcp;
2654 u8 last_flags;
2655
2656 idx = tx->req & tx->mask;
2657
2658 last_flags = src->flags;
2659 src->flags = 0;
2660 mb();
2661 dst = dstp = &tx->lanai[idx];
2662 srcp = src;
2663
2664 if ((idx + cnt) < tx->mask) {
2665 for (i = 0; i < (cnt - 1); i += 2) {
2666 myri10ge_pio_copy(dstp, srcp, 2 * sizeof(*src));
2667 mb(); /* force write every 32 bytes */
2668 srcp += 2;
2669 dstp += 2;
2670 }
2671 } else {
2672 /* submit all but the first request, and ensure
2673 * that it is submitted below */
2674 myri10ge_submit_req_backwards(tx, src, cnt);
2675 i = 0;
2676 }
2677 if (i < cnt) {
2678 /* submit the first request */
2679 myri10ge_pio_copy(dstp, srcp, sizeof(*src));
2680 mb(); /* barrier before setting valid flag */
2681 }
2682
2683 /* re-write the last 32-bits with the valid flags */
2684 src->flags = last_flags;
40f6cff5 2685 put_be32(*((__be32 *) src + 3), (__be32 __iomem *) dst + 3);
0da34b6d
BG
2686 tx->req += cnt;
2687 mb();
2688}
2689
0da34b6d
BG
2690/*
2691 * Transmit a packet. We need to split the packet so that a single
b53bef84 2692 * segment does not cross myri10ge->tx_boundary, so this makes segment
0da34b6d
BG
2693 * counting tricky. So rather than try to count segments up front, we
2694 * just give up if there are too few segments to hold a reasonably
2695 * fragmented packet currently available. If we run
2696 * out of segments while preparing a packet for DMA, we just linearize
2697 * it and try again.
2698 */
2699
61357325
SH
2700static netdev_tx_t myri10ge_xmit(struct sk_buff *skb,
2701 struct net_device *dev)
0da34b6d
BG
2702{
2703 struct myri10ge_priv *mgp = netdev_priv(dev);
b53bef84 2704 struct myri10ge_slice_state *ss;
0da34b6d 2705 struct mcp_kreq_ether_send *req;
b53bef84 2706 struct myri10ge_tx_buf *tx;
0da34b6d 2707 struct skb_frag_struct *frag;
236bb5e6 2708 struct netdev_queue *netdev_queue;
0da34b6d 2709 dma_addr_t bus;
40f6cff5
AV
2710 u32 low;
2711 __be32 high_swapped;
0da34b6d
BG
2712 unsigned int len;
2713 int idx, last_idx, avail, frag_cnt, frag_idx, count, mss, max_segments;
236bb5e6 2714 u16 pseudo_hdr_offset, cksum_offset, queue;
0da34b6d
BG
2715 int cum_len, seglen, boundary, rdma_count;
2716 u8 flags, odd_flag;
2717
236bb5e6 2718 queue = skb_get_queue_mapping(skb);
236bb5e6
BG
2719 ss = &mgp->ss[queue];
2720 netdev_queue = netdev_get_tx_queue(mgp->dev, queue);
b53bef84 2721 tx = &ss->tx;
236bb5e6 2722
0da34b6d
BG
2723again:
2724 req = tx->req_list;
2725 avail = tx->mask - 1 - (tx->req - tx->done);
2726
2727 mss = 0;
2728 max_segments = MXGEFW_MAX_SEND_DESC;
2729
917690cd 2730 if (skb_is_gso(skb)) {
7967168c 2731 mss = skb_shinfo(skb)->gso_size;
917690cd 2732 max_segments = MYRI10GE_MAX_SEND_DESC_TSO;
0da34b6d 2733 }
0da34b6d
BG
2734
2735 if ((unlikely(avail < max_segments))) {
2736 /* we are out of transmit resources */
b53bef84 2737 tx->stop_queue++;
236bb5e6 2738 netif_tx_stop_queue(netdev_queue);
5b548140 2739 return NETDEV_TX_BUSY;
0da34b6d
BG
2740 }
2741
2742 /* Setup checksum offloading, if needed */
2743 cksum_offset = 0;
2744 pseudo_hdr_offset = 0;
2745 odd_flag = 0;
2746 flags = (MXGEFW_FLAGS_NO_TSO | MXGEFW_FLAGS_FIRST);
84fa7933 2747 if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
0d0b1672 2748 cksum_offset = skb_checksum_start_offset(skb);
ff1dcadb 2749 pseudo_hdr_offset = cksum_offset + skb->csum_offset;
0da34b6d
BG
2750 /* If the headers are excessively large, then we must
2751 * fall back to a software checksum */
4f93fde0
BG
2752 if (unlikely(!mss && (cksum_offset > 255 ||
2753 pseudo_hdr_offset > 127))) {
84fa7933 2754 if (skb_checksum_help(skb))
0da34b6d
BG
2755 goto drop;
2756 cksum_offset = 0;
2757 pseudo_hdr_offset = 0;
2758 } else {
0da34b6d
BG
2759 odd_flag = MXGEFW_FLAGS_ALIGN_ODD;
2760 flags |= MXGEFW_FLAGS_CKSUM;
2761 }
2762 }
2763
2764 cum_len = 0;
2765
0da34b6d
BG
2766 if (mss) { /* TSO */
2767 /* this removes any CKSUM flag from before */
2768 flags = (MXGEFW_FLAGS_TSO_HDR | MXGEFW_FLAGS_FIRST);
2769
2770 /* negative cum_len signifies to the
2771 * send loop that we are still in the
2772 * header portion of the TSO packet.
4f93fde0 2773 * TSO header can be at most 1KB long */
ab6a5bb6 2774 cum_len = -(skb_transport_offset(skb) + tcp_hdrlen(skb));
0da34b6d 2775
4f93fde0
BG
2776 /* for IPv6 TSO, the checksum offset stores the
2777 * TCP header length, to save the firmware from
2778 * the need to parse the headers */
2779 if (skb_is_gso_v6(skb)) {
2780 cksum_offset = tcp_hdrlen(skb);
2781 /* Can only handle headers <= max_tso6 long */
2782 if (unlikely(-cum_len > mgp->max_tso6))
2783 return myri10ge_sw_tso(skb, dev);
2784 }
0da34b6d
BG
2785 /* for TSO, pseudo_hdr_offset holds mss.
2786 * The firmware figures out where to put
2787 * the checksum by parsing the header. */
40f6cff5 2788 pseudo_hdr_offset = mss;
0da34b6d 2789 } else
0da34b6d
BG
2790 /* Mark small packets, and pad out tiny packets */
2791 if (skb->len <= MXGEFW_SEND_SMALL_SIZE) {
2792 flags |= MXGEFW_FLAGS_SMALL;
2793
2794 /* pad frames to at least ETH_ZLEN bytes */
2795 if (unlikely(skb->len < ETH_ZLEN)) {
5b057c6b 2796 if (skb_padto(skb, ETH_ZLEN)) {
0da34b6d
BG
2797 /* The packet is gone, so we must
2798 * return 0 */
b53bef84 2799 ss->stats.tx_dropped += 1;
6ed10654 2800 return NETDEV_TX_OK;
0da34b6d
BG
2801 }
2802 /* adjust the len to account for the zero pad
2803 * so that the nic can know how long it is */
2804 skb->len = ETH_ZLEN;
2805 }
2806 }
2807
2808 /* map the skb for DMA */
e743d313 2809 len = skb_headlen(skb);
0da34b6d
BG
2810 idx = tx->req & tx->mask;
2811 tx->info[idx].skb = skb;
2812 bus = pci_map_single(mgp->pdev, skb->data, len, PCI_DMA_TODEVICE);
c755b4b6
FT
2813 dma_unmap_addr_set(&tx->info[idx], bus, bus);
2814 dma_unmap_len_set(&tx->info[idx], len, len);
0da34b6d
BG
2815
2816 frag_cnt = skb_shinfo(skb)->nr_frags;
2817 frag_idx = 0;
2818 count = 0;
2819 rdma_count = 0;
2820
2821 /* "rdma_count" is the number of RDMAs belonging to the
2822 * current packet BEFORE the current send request. For
2823 * non-TSO packets, this is equal to "count".
2824 * For TSO packets, rdma_count needs to be reset
2825 * to 0 after a segment cut.
2826 *
2827 * The rdma_count field of the send request is
2828 * the number of RDMAs of the packet starting at
2829 * that request. For TSO send requests with one ore more cuts
2830 * in the middle, this is the number of RDMAs starting
2831 * after the last cut in the request. All previous
2832 * segments before the last cut implicitly have 1 RDMA.
2833 *
2834 * Since the number of RDMAs is not known beforehand,
2835 * it must be filled-in retroactively - after each
2836 * segmentation cut or at the end of the entire packet.
2837 */
2838
2839 while (1) {
2840 /* Break the SKB or Fragment up into pieces which
b53bef84 2841 * do not cross mgp->tx_boundary */
0da34b6d
BG
2842 low = MYRI10GE_LOWPART_TO_U32(bus);
2843 high_swapped = htonl(MYRI10GE_HIGHPART_TO_U32(bus));
2844 while (len) {
2845 u8 flags_next;
2846 int cum_len_next;
2847
2848 if (unlikely(count == max_segments))
2849 goto abort_linearize;
2850
b53bef84
BG
2851 boundary =
2852 (low + mgp->tx_boundary) & ~(mgp->tx_boundary - 1);
0da34b6d
BG
2853 seglen = boundary - low;
2854 if (seglen > len)
2855 seglen = len;
2856 flags_next = flags & ~MXGEFW_FLAGS_FIRST;
2857 cum_len_next = cum_len + seglen;
0da34b6d
BG
2858 if (mss) { /* TSO */
2859 (req - rdma_count)->rdma_count = rdma_count + 1;
2860
2861 if (likely(cum_len >= 0)) { /* payload */
2862 int next_is_first, chop;
2863
2864 chop = (cum_len_next > mss);
2865 cum_len_next = cum_len_next % mss;
2866 next_is_first = (cum_len_next == 0);
2867 flags |= chop * MXGEFW_FLAGS_TSO_CHOP;
2868 flags_next |= next_is_first *
2869 MXGEFW_FLAGS_FIRST;
2870 rdma_count |= -(chop | next_is_first);
2871 rdma_count += chop & !next_is_first;
2872 } else if (likely(cum_len_next >= 0)) { /* header ends */
2873 int small;
2874
2875 rdma_count = -1;
2876 cum_len_next = 0;
2877 seglen = -cum_len;
2878 small = (mss <= MXGEFW_SEND_SMALL_SIZE);
2879 flags_next = MXGEFW_FLAGS_TSO_PLD |
2880 MXGEFW_FLAGS_FIRST |
2881 (small * MXGEFW_FLAGS_SMALL);
2882 }
2883 }
0da34b6d
BG
2884 req->addr_high = high_swapped;
2885 req->addr_low = htonl(low);
40f6cff5 2886 req->pseudo_hdr_offset = htons(pseudo_hdr_offset);
0da34b6d
BG
2887 req->pad = 0; /* complete solid 16-byte block; does this matter? */
2888 req->rdma_count = 1;
2889 req->length = htons(seglen);
2890 req->cksum_offset = cksum_offset;
2891 req->flags = flags | ((cum_len & 1) * odd_flag);
2892
2893 low += seglen;
2894 len -= seglen;
2895 cum_len = cum_len_next;
2896 flags = flags_next;
2897 req++;
2898 count++;
2899 rdma_count++;
4f93fde0
BG
2900 if (cksum_offset != 0 && !(mss && skb_is_gso_v6(skb))) {
2901 if (unlikely(cksum_offset > seglen))
2902 cksum_offset -= seglen;
2903 else
2904 cksum_offset = 0;
2905 }
0da34b6d
BG
2906 }
2907 if (frag_idx == frag_cnt)
2908 break;
2909
2910 /* map next fragment for DMA */
2911 idx = (count + tx->req) & tx->mask;
2912 frag = &skb_shinfo(skb)->frags[frag_idx];
2913 frag_idx++;
2914 len = frag->size;
2915 bus = pci_map_page(mgp->pdev, frag->page, frag->page_offset,
2916 len, PCI_DMA_TODEVICE);
c755b4b6
FT
2917 dma_unmap_addr_set(&tx->info[idx], bus, bus);
2918 dma_unmap_len_set(&tx->info[idx], len, len);
0da34b6d
BG
2919 }
2920
2921 (req - rdma_count)->rdma_count = rdma_count;
0da34b6d
BG
2922 if (mss)
2923 do {
2924 req--;
2925 req->flags |= MXGEFW_FLAGS_TSO_LAST;
2926 } while (!(req->flags & (MXGEFW_FLAGS_TSO_CHOP |
2927 MXGEFW_FLAGS_FIRST)));
0da34b6d
BG
2928 idx = ((count - 1) + tx->req) & tx->mask;
2929 tx->info[idx].last = 1;
e454e7e2 2930 myri10ge_submit_req(tx, tx->req_list, count);
236bb5e6
BG
2931 /* if using multiple tx queues, make sure NIC polls the
2932 * current slice */
2933 if ((mgp->dev->real_num_tx_queues > 1) && tx->queue_active == 0) {
2934 tx->queue_active = 1;
2935 put_be32(htonl(1), tx->send_go);
8c2f5fa5 2936 mb();
6824a105 2937 mmiowb();
236bb5e6 2938 }
0da34b6d
BG
2939 tx->pkt_start++;
2940 if ((avail - count) < MXGEFW_MAX_SEND_DESC) {
b53bef84 2941 tx->stop_queue++;
236bb5e6 2942 netif_tx_stop_queue(netdev_queue);
0da34b6d 2943 }
6ed10654 2944 return NETDEV_TX_OK;
0da34b6d
BG
2945
2946abort_linearize:
2947 /* Free any DMA resources we've alloced and clear out the skb
2948 * slot so as to not trip up assertions, and to avoid a
2949 * double-free if linearizing fails */
2950
2951 last_idx = (idx + 1) & tx->mask;
2952 idx = tx->req & tx->mask;
2953 tx->info[idx].skb = NULL;
2954 do {
c755b4b6 2955 len = dma_unmap_len(&tx->info[idx], len);
0da34b6d
BG
2956 if (len) {
2957 if (tx->info[idx].skb != NULL)
2958 pci_unmap_single(mgp->pdev,
c755b4b6 2959 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
2960 bus), len,
2961 PCI_DMA_TODEVICE);
2962 else
2963 pci_unmap_page(mgp->pdev,
c755b4b6 2964 dma_unmap_addr(&tx->info[idx],
0da34b6d
BG
2965 bus), len,
2966 PCI_DMA_TODEVICE);
c755b4b6 2967 dma_unmap_len_set(&tx->info[idx], len, 0);
0da34b6d
BG
2968 tx->info[idx].skb = NULL;
2969 }
2970 idx = (idx + 1) & tx->mask;
2971 } while (idx != last_idx);
89114afd 2972 if (skb_is_gso(skb)) {
78ca90ea 2973 netdev_err(mgp->dev, "TSO but wanted to linearize?!?!?\n");
0da34b6d
BG
2974 goto drop;
2975 }
2976
bec0e859 2977 if (skb_linearize(skb))
0da34b6d
BG
2978 goto drop;
2979
b53bef84 2980 tx->linearized++;
0da34b6d
BG
2981 goto again;
2982
2983drop:
2984 dev_kfree_skb_any(skb);
b53bef84 2985 ss->stats.tx_dropped += 1;
6ed10654 2986 return NETDEV_TX_OK;
0da34b6d
BG
2987
2988}
2989
61357325
SH
2990static netdev_tx_t myri10ge_sw_tso(struct sk_buff *skb,
2991 struct net_device *dev)
4f93fde0
BG
2992{
2993 struct sk_buff *segs, *curr;
b53bef84 2994 struct myri10ge_priv *mgp = netdev_priv(dev);
d6279c88 2995 struct myri10ge_slice_state *ss;
61357325 2996 netdev_tx_t status;
4f93fde0
BG
2997
2998 segs = skb_gso_segment(skb, dev->features & ~NETIF_F_TSO6);
801678c5 2999 if (IS_ERR(segs))
4f93fde0
BG
3000 goto drop;
3001
3002 while (segs) {
3003 curr = segs;
3004 segs = segs->next;
3005 curr->next = NULL;
3006 status = myri10ge_xmit(curr, dev);
3007 if (status != 0) {
3008 dev_kfree_skb_any(curr);
3009 if (segs != NULL) {
3010 curr = segs;
3011 segs = segs->next;
3012 curr->next = NULL;
3013 dev_kfree_skb_any(segs);
3014 }
3015 goto drop;
3016 }
3017 }
3018 dev_kfree_skb_any(skb);
ec634fe3 3019 return NETDEV_TX_OK;
4f93fde0
BG
3020
3021drop:
d6279c88 3022 ss = &mgp->ss[skb_get_queue_mapping(skb)];
4f93fde0 3023 dev_kfree_skb_any(skb);
d6279c88 3024 ss->stats.tx_dropped += 1;
ec634fe3 3025 return NETDEV_TX_OK;
4f93fde0
BG
3026}
3027
0da34b6d
BG
3028static struct net_device_stats *myri10ge_get_stats(struct net_device *dev)
3029{
3030 struct myri10ge_priv *mgp = netdev_priv(dev);
0dcffac1 3031 struct myri10ge_slice_netstats *slice_stats;
6dc34941 3032 struct net_device_stats *stats = &dev->stats;
0dcffac1
BG
3033 int i;
3034
59081825 3035 spin_lock(&mgp->stats_lock);
0dcffac1
BG
3036 memset(stats, 0, sizeof(*stats));
3037 for (i = 0; i < mgp->num_slices; i++) {
3038 slice_stats = &mgp->ss[i].stats;
3039 stats->rx_packets += slice_stats->rx_packets;
3040 stats->tx_packets += slice_stats->tx_packets;
3041 stats->rx_bytes += slice_stats->rx_bytes;
3042 stats->tx_bytes += slice_stats->tx_bytes;
3043 stats->rx_dropped += slice_stats->rx_dropped;
3044 stats->tx_dropped += slice_stats->tx_dropped;
3045 }
59081825 3046 spin_unlock(&mgp->stats_lock);
0dcffac1 3047 return stats;
0da34b6d
BG
3048}
3049
3050static void myri10ge_set_multicast_list(struct net_device *dev)
3051{
b53bef84 3052 struct myri10ge_priv *mgp = netdev_priv(dev);
85a7ea1b 3053 struct myri10ge_cmd cmd;
22bedad3 3054 struct netdev_hw_addr *ha;
6250223e 3055 __be32 data[2] = { 0, 0 };
85a7ea1b
BG
3056 int err;
3057
0da34b6d
BG
3058 /* can be called from atomic contexts,
3059 * pass 1 to force atomicity in myri10ge_send_cmd() */
85a7ea1b
BG
3060 myri10ge_change_promisc(mgp, dev->flags & IFF_PROMISC, 1);
3061
3062 /* This firmware is known to not support multicast */
2f76216f 3063 if (!mgp->fw_multicast_support)
85a7ea1b
BG
3064 return;
3065
3066 /* Disable multicast filtering */
3067
3068 err = myri10ge_send_cmd(mgp, MXGEFW_ENABLE_ALLMULTI, &cmd, 1);
3069 if (err != 0) {
78ca90ea
JP
3070 netdev_err(dev, "Failed MXGEFW_ENABLE_ALLMULTI, error status: %d\n",
3071 err);
85a7ea1b
BG
3072 goto abort;
3073 }
3074
2f76216f 3075 if ((dev->flags & IFF_ALLMULTI) || mgp->adopted_rx_filter_bug) {
85a7ea1b
BG
3076 /* request to disable multicast filtering, so quit here */
3077 return;
3078 }
3079
3080 /* Flush the filters */
3081
3082 err = myri10ge_send_cmd(mgp, MXGEFW_LEAVE_ALL_MULTICAST_GROUPS,
3083 &cmd, 1);
3084 if (err != 0) {
78ca90ea
JP
3085 netdev_err(dev, "Failed MXGEFW_LEAVE_ALL_MULTICAST_GROUPS, error status: %d\n",
3086 err);
85a7ea1b
BG
3087 goto abort;
3088 }
3089
3090 /* Walk the multicast list, and add each address */
22bedad3
JP
3091 netdev_for_each_mc_addr(ha, dev) {
3092 memcpy(data, &ha->addr, 6);
40f6cff5
AV
3093 cmd.data0 = ntohl(data[0]);
3094 cmd.data1 = ntohl(data[1]);
85a7ea1b
BG
3095 err = myri10ge_send_cmd(mgp, MXGEFW_JOIN_MULTICAST_GROUP,
3096 &cmd, 1);
3097
3098 if (err != 0) {
78ca90ea 3099 netdev_err(dev, "Failed MXGEFW_JOIN_MULTICAST_GROUP, error status:%d %pM\n",
22bedad3 3100 err, ha->addr);
85a7ea1b
BG
3101 goto abort;
3102 }
3103 }
3104 /* Enable multicast filtering */
3105 err = myri10ge_send_cmd(mgp, MXGEFW_DISABLE_ALLMULTI, &cmd, 1);
3106 if (err != 0) {
78ca90ea
JP
3107 netdev_err(dev, "Failed MXGEFW_DISABLE_ALLMULTI, error status: %d\n",
3108 err);
85a7ea1b
BG
3109 goto abort;
3110 }
3111
3112 return;
3113
3114abort:
3115 return;
0da34b6d
BG
3116}
3117
3118static int myri10ge_set_mac_address(struct net_device *dev, void *addr)
3119{
3120 struct sockaddr *sa = addr;
3121 struct myri10ge_priv *mgp = netdev_priv(dev);
3122 int status;
3123
3124 if (!is_valid_ether_addr(sa->sa_data))
3125 return -EADDRNOTAVAIL;
3126
3127 status = myri10ge_update_mac_address(mgp, sa->sa_data);
3128 if (status != 0) {
78ca90ea
JP
3129 netdev_err(dev, "changing mac address failed with %d\n",
3130 status);
0da34b6d
BG
3131 return status;
3132 }
3133
3134 /* change the dev structure */
3135 memcpy(dev->dev_addr, sa->sa_data, 6);
3136 return 0;
3137}
3138
3139static int myri10ge_change_mtu(struct net_device *dev, int new_mtu)
3140{
3141 struct myri10ge_priv *mgp = netdev_priv(dev);
3142 int error = 0;
3143
3144 if ((new_mtu < 68) || (ETH_HLEN + new_mtu > MYRI10GE_MAX_ETHER_MTU)) {
78ca90ea 3145 netdev_err(dev, "new mtu (%d) is not valid\n", new_mtu);
0da34b6d
BG
3146 return -EINVAL;
3147 }
78ca90ea 3148 netdev_info(dev, "changing mtu from %d to %d\n", dev->mtu, new_mtu);
0da34b6d
BG
3149 if (mgp->running) {
3150 /* if we change the mtu on an active device, we must
3151 * reset the device so the firmware sees the change */
3152 myri10ge_close(dev);
3153 dev->mtu = new_mtu;
3154 myri10ge_open(dev);
3155 } else
3156 dev->mtu = new_mtu;
3157
3158 return error;
3159}
3160
3161/*
3162 * Enable ECRC to align PCI-E Completion packets on an 8-byte boundary.
3163 * Only do it if the bridge is a root port since we don't want to disturb
3164 * any other device, except if forced with myri10ge_ecrc_enable > 1.
3165 */
3166
0da34b6d
BG
3167static void myri10ge_enable_ecrc(struct myri10ge_priv *mgp)
3168{
3169 struct pci_dev *bridge = mgp->pdev->bus->self;
3170 struct device *dev = &mgp->pdev->dev;
3171 unsigned cap;
3172 unsigned err_cap;
3173 u16 val;
3174 u8 ext_type;
3175 int ret;
3176
3177 if (!myri10ge_ecrc_enable || !bridge)
3178 return;
3179
3180 /* check that the bridge is a root port */
3181 cap = pci_find_capability(bridge, PCI_CAP_ID_EXP);
3182 pci_read_config_word(bridge, cap + PCI_CAP_FLAGS, &val);
3183 ext_type = (val & PCI_EXP_FLAGS_TYPE) >> 4;
3184 if (ext_type != PCI_EXP_TYPE_ROOT_PORT) {
3185 if (myri10ge_ecrc_enable > 1) {
eca3fd83 3186 struct pci_dev *prev_bridge, *old_bridge = bridge;
0da34b6d
BG
3187
3188 /* Walk the hierarchy up to the root port
3189 * where ECRC has to be enabled */
3190 do {
eca3fd83 3191 prev_bridge = bridge;
0da34b6d 3192 bridge = bridge->bus->self;
eca3fd83 3193 if (!bridge || prev_bridge == bridge) {
0da34b6d
BG
3194 dev_err(dev,
3195 "Failed to find root port"
3196 " to force ECRC\n");
3197 return;
3198 }
3199 cap =
3200 pci_find_capability(bridge, PCI_CAP_ID_EXP);
3201 pci_read_config_word(bridge,
3202 cap + PCI_CAP_FLAGS, &val);
3203 ext_type = (val & PCI_EXP_FLAGS_TYPE) >> 4;
3204 } while (ext_type != PCI_EXP_TYPE_ROOT_PORT);
3205
3206 dev_info(dev,
3207 "Forcing ECRC on non-root port %s"
3208 " (enabling on root port %s)\n",
3209 pci_name(old_bridge), pci_name(bridge));
3210 } else {
3211 dev_err(dev,
3212 "Not enabling ECRC on non-root port %s\n",
3213 pci_name(bridge));
3214 return;
3215 }
3216 }
3217
3218 cap = pci_find_ext_capability(bridge, PCI_EXT_CAP_ID_ERR);
0da34b6d
BG
3219 if (!cap)
3220 return;
3221
3222 ret = pci_read_config_dword(bridge, cap + PCI_ERR_CAP, &err_cap);
3223 if (ret) {
3224 dev_err(dev, "failed reading ext-conf-space of %s\n",
3225 pci_name(bridge));
3226 dev_err(dev, "\t pci=nommconf in use? "
3227 "or buggy/incomplete/absent ACPI MCFG attr?\n");
3228 return;
3229 }
3230 if (!(err_cap & PCI_ERR_CAP_ECRC_GENC))
3231 return;
3232
3233 err_cap |= PCI_ERR_CAP_ECRC_GENE;
3234 pci_write_config_dword(bridge, cap + PCI_ERR_CAP, err_cap);
3235 dev_info(dev, "Enabled ECRC on upstream bridge %s\n", pci_name(bridge));
0da34b6d
BG
3236}
3237
3238/*
3239 * The Lanai Z8E PCI-E interface achieves higher Read-DMA throughput
3240 * when the PCI-E Completion packets are aligned on an 8-byte
3241 * boundary. Some PCI-E chip sets always align Completion packets; on
3242 * the ones that do not, the alignment can be enforced by enabling
3243 * ECRC generation (if supported).
3244 *
3245 * When PCI-E Completion packets are not aligned, it is actually more
3246 * efficient to limit Read-DMA transactions to 2KB, rather than 4KB.
3247 *
3248 * If the driver can neither enable ECRC nor verify that it has
3249 * already been enabled, then it must use a firmware image which works
0dcffac1 3250 * around unaligned completion packets (myri10ge_rss_ethp_z8e.dat), and it
0da34b6d 3251 * should also ensure that it never gives the device a Read-DMA which is
b53bef84 3252 * larger than 2KB by setting the tx_boundary to 2KB. If ECRC is
0dcffac1 3253 * enabled, then the driver should use the aligned (myri10ge_rss_eth_z8e.dat)
b53bef84 3254 * firmware image, and set tx_boundary to 4KB.
0da34b6d
BG
3255 */
3256
5443e9ea 3257static void myri10ge_firmware_probe(struct myri10ge_priv *mgp)
0da34b6d 3258{
5443e9ea
BG
3259 struct pci_dev *pdev = mgp->pdev;
3260 struct device *dev = &pdev->dev;
302d242c 3261 int status;
0da34b6d 3262
b53bef84 3263 mgp->tx_boundary = 4096;
5443e9ea
BG
3264 /*
3265 * Verify the max read request size was set to 4KB
3266 * before trying the test with 4KB.
3267 */
302d242c
BG
3268 status = pcie_get_readrq(pdev);
3269 if (status < 0) {
5443e9ea
BG
3270 dev_err(dev, "Couldn't read max read req size: %d\n", status);
3271 goto abort;
3272 }
302d242c
BG
3273 if (status != 4096) {
3274 dev_warn(dev, "Max Read Request size != 4096 (%d)\n", status);
b53bef84 3275 mgp->tx_boundary = 2048;
5443e9ea
BG
3276 }
3277 /*
3278 * load the optimized firmware (which assumes aligned PCIe
3279 * completions) in order to see if it works on this host.
3280 */
7d351035 3281 set_fw_name(mgp, myri10ge_fw_aligned, false);
0dcffac1 3282 status = myri10ge_load_firmware(mgp, 1);
5443e9ea
BG
3283 if (status != 0) {
3284 goto abort;
3285 }
3286
3287 /*
3288 * Enable ECRC if possible
3289 */
3290 myri10ge_enable_ecrc(mgp);
3291
3292 /*
3293 * Run a DMA test which watches for unaligned completions and
3294 * aborts on the first one seen.
3295 */
3296
3297 status = myri10ge_dma_test(mgp, MXGEFW_CMD_UNALIGNED_TEST);
3298 if (status == 0)
3299 return; /* keep the aligned firmware */
3300
3301 if (status != -E2BIG)
3302 dev_warn(dev, "DMA test failed: %d\n", status);
3303 if (status == -ENOSYS)
3304 dev_warn(dev, "Falling back to ethp! "
3305 "Please install up to date fw\n");
3306abort:
3307 /* fall back to using the unaligned firmware */
b53bef84 3308 mgp->tx_boundary = 2048;
7d351035 3309 set_fw_name(mgp, myri10ge_fw_unaligned, false);
0da34b6d 3310
5443e9ea
BG
3311}
3312
3313static void myri10ge_select_firmware(struct myri10ge_priv *mgp)
3314{
2d90b0aa
BG
3315 int overridden = 0;
3316
0da34b6d 3317 if (myri10ge_force_firmware == 0) {
ce7f9368
BG
3318 int link_width, exp_cap;
3319 u16 lnk;
3320
3321 exp_cap = pci_find_capability(mgp->pdev, PCI_CAP_ID_EXP);
3322 pci_read_config_word(mgp->pdev, exp_cap + PCI_EXP_LNKSTA, &lnk);
3323 link_width = (lnk >> 4) & 0x3f;
3324
ce7f9368
BG
3325 /* Check to see if Link is less than 8 or if the
3326 * upstream bridge is known to provide aligned
3327 * completions */
3328 if (link_width < 8) {
3329 dev_info(&mgp->pdev->dev, "PCIE x%d Link\n",
3330 link_width);
b53bef84 3331 mgp->tx_boundary = 4096;
7d351035 3332 set_fw_name(mgp, myri10ge_fw_aligned, false);
5443e9ea
BG
3333 } else {
3334 myri10ge_firmware_probe(mgp);
0da34b6d
BG
3335 }
3336 } else {
3337 if (myri10ge_force_firmware == 1) {
3338 dev_info(&mgp->pdev->dev,
3339 "Assuming aligned completions (forced)\n");
b53bef84 3340 mgp->tx_boundary = 4096;
7d351035 3341 set_fw_name(mgp, myri10ge_fw_aligned, false);
0da34b6d
BG
3342 } else {
3343 dev_info(&mgp->pdev->dev,
3344 "Assuming unaligned completions (forced)\n");
b53bef84 3345 mgp->tx_boundary = 2048;
7d351035 3346 set_fw_name(mgp, myri10ge_fw_unaligned, false);
0da34b6d
BG
3347 }
3348 }
7d351035
RR
3349
3350 kparam_block_sysfs_write(myri10ge_fw_name);
0da34b6d 3351 if (myri10ge_fw_name != NULL) {
7d351035
RR
3352 char *fw_name = kstrdup(myri10ge_fw_name, GFP_KERNEL);
3353 if (fw_name) {
3354 overridden = 1;
3355 set_fw_name(mgp, fw_name, true);
3356 }
0da34b6d 3357 }
7d351035
RR
3358 kparam_unblock_sysfs_write(myri10ge_fw_name);
3359
2d90b0aa
BG
3360 if (mgp->board_number < MYRI10GE_MAX_BOARDS &&
3361 myri10ge_fw_names[mgp->board_number] != NULL &&
3362 strlen(myri10ge_fw_names[mgp->board_number])) {
7d351035 3363 set_fw_name(mgp, myri10ge_fw_names[mgp->board_number], false);
2d90b0aa
BG
3364 overridden = 1;
3365 }
3366 if (overridden)
3367 dev_info(&mgp->pdev->dev, "overriding firmware to %s\n",
3368 mgp->fw_name);
0da34b6d
BG
3369}
3370
0da34b6d 3371#ifdef CONFIG_PM
0da34b6d
BG
3372static int myri10ge_suspend(struct pci_dev *pdev, pm_message_t state)
3373{
3374 struct myri10ge_priv *mgp;
3375 struct net_device *netdev;
3376
3377 mgp = pci_get_drvdata(pdev);
3378 if (mgp == NULL)
3379 return -EINVAL;
3380 netdev = mgp->dev;
3381
3382 netif_device_detach(netdev);
3383 if (netif_running(netdev)) {
78ca90ea 3384 netdev_info(netdev, "closing\n");
0da34b6d
BG
3385 rtnl_lock();
3386 myri10ge_close(netdev);
3387 rtnl_unlock();
3388 }
3389 myri10ge_dummy_rdma(mgp, 0);
83f6e152 3390 pci_save_state(pdev);
0da34b6d 3391 pci_disable_device(pdev);
1a63e846
BG
3392
3393 return pci_set_power_state(pdev, pci_choose_state(pdev, state));
0da34b6d
BG
3394}
3395
3396static int myri10ge_resume(struct pci_dev *pdev)
3397{
3398 struct myri10ge_priv *mgp;
3399 struct net_device *netdev;
3400 int status;
3401 u16 vendor;
3402
3403 mgp = pci_get_drvdata(pdev);
3404 if (mgp == NULL)
3405 return -EINVAL;
3406 netdev = mgp->dev;
3407 pci_set_power_state(pdev, 0); /* zeros conf space as a side effect */
3408 msleep(5); /* give card time to respond */
3409 pci_read_config_word(mgp->pdev, PCI_VENDOR_ID, &vendor);
3410 if (vendor == 0xffff) {
78ca90ea 3411 netdev_err(mgp->dev, "device disappeared!\n");
0da34b6d
BG
3412 return -EIO;
3413 }
83f6e152 3414
1d3c16a8 3415 pci_restore_state(pdev);
4c2248cc
BG
3416
3417 status = pci_enable_device(pdev);
1a63e846 3418 if (status) {
4c2248cc 3419 dev_err(&pdev->dev, "failed to enable device\n");
1a63e846 3420 return status;
4c2248cc
BG
3421 }
3422
0da34b6d
BG
3423 pci_set_master(pdev);
3424
0da34b6d 3425 myri10ge_reset(mgp);
013b68bf 3426 myri10ge_dummy_rdma(mgp, 1);
0da34b6d
BG
3427
3428 /* Save configuration space to be restored if the
3429 * nic resets due to a parity error */
83f6e152 3430 pci_save_state(pdev);
0da34b6d
BG
3431
3432 if (netif_running(netdev)) {
3433 rtnl_lock();
df30a740 3434 status = myri10ge_open(netdev);
0da34b6d 3435 rtnl_unlock();
df30a740
BG
3436 if (status != 0)
3437 goto abort_with_enabled;
3438
0da34b6d
BG
3439 }
3440 netif_device_attach(netdev);
3441
3442 return 0;
3443
4c2248cc
BG
3444abort_with_enabled:
3445 pci_disable_device(pdev);
0da34b6d
BG
3446 return -EIO;
3447
3448}
0da34b6d
BG
3449#endif /* CONFIG_PM */
3450
3451static u32 myri10ge_read_reboot(struct myri10ge_priv *mgp)
3452{
3453 struct pci_dev *pdev = mgp->pdev;
3454 int vs = mgp->vendor_specific_offset;
3455 u32 reboot;
3456
3457 /*enter read32 mode */
3458 pci_write_config_byte(pdev, vs + 0x10, 0x3);
3459
3460 /*read REBOOT_STATUS (0xfffffff0) */
3461 pci_write_config_dword(pdev, vs + 0x18, 0xfffffff0);
3462 pci_read_config_dword(pdev, vs + 0x14, &reboot);
3463 return reboot;
3464}
3465
3466/*
3467 * This watchdog is used to check whether the board has suffered
3468 * from a parity error and needs to be recovered.
3469 */
c4028958 3470static void myri10ge_watchdog(struct work_struct *work)
0da34b6d 3471{
c4028958 3472 struct myri10ge_priv *mgp =
6250223e 3473 container_of(work, struct myri10ge_priv, watchdog_work);
b53bef84 3474 struct myri10ge_tx_buf *tx;
0da34b6d 3475 u32 reboot;
d0234215 3476 int status, rebooted;
0dcffac1 3477 int i;
0da34b6d
BG
3478 u16 cmd, vendor;
3479
3480 mgp->watchdog_resets++;
3481 pci_read_config_word(mgp->pdev, PCI_COMMAND, &cmd);
d0234215 3482 rebooted = 0;
0da34b6d
BG
3483 if ((cmd & PCI_COMMAND_MASTER) == 0) {
3484 /* Bus master DMA disabled? Check to see
3485 * if the card rebooted due to a parity error
3486 * For now, just report it */
3487 reboot = myri10ge_read_reboot(mgp);
78ca90ea
JP
3488 netdev_err(mgp->dev, "NIC rebooted (0x%x),%s resetting\n",
3489 reboot,
3490 myri10ge_reset_recover ? "" : " not");
f181137f
BG
3491 if (myri10ge_reset_recover == 0)
3492 return;
d0234215
BG
3493 rtnl_lock();
3494 mgp->rebooted = 1;
3495 rebooted = 1;
3496 myri10ge_close(mgp->dev);
f181137f 3497 myri10ge_reset_recover--;
d0234215 3498 mgp->rebooted = 0;
0da34b6d
BG
3499 /*
3500 * A rebooted nic will come back with config space as
3501 * it was after power was applied to PCIe bus.
3502 * Attempt to restore config space which was saved
3503 * when the driver was loaded, or the last time the
3504 * nic was resumed from power saving mode.
3505 */
83f6e152 3506 pci_restore_state(mgp->pdev);
7adda30c
BG
3507
3508 /* save state again for accounting reasons */
83f6e152 3509 pci_save_state(mgp->pdev);
7adda30c 3510
0da34b6d
BG
3511 } else {
3512 /* if we get back -1's from our slot, perhaps somebody
3513 * powered off our card. Don't try to reset it in
3514 * this case */
3515 if (cmd == 0xffff) {
3516 pci_read_config_word(mgp->pdev, PCI_VENDOR_ID, &vendor);
3517 if (vendor == 0xffff) {
78ca90ea 3518 netdev_err(mgp->dev, "device disappeared!\n");
0da34b6d
BG
3519 return;
3520 }
3521 }
3522 /* Perhaps it is a software error. Try to reset */
3523
78ca90ea 3524 netdev_err(mgp->dev, "device timeout, resetting\n");
0dcffac1
BG
3525 for (i = 0; i < mgp->num_slices; i++) {
3526 tx = &mgp->ss[i].tx;
78ca90ea
JP
3527 netdev_err(mgp->dev, "(%d): %d %d %d %d %d %d\n",
3528 i, tx->queue_active, tx->req,
3529 tx->done, tx->pkt_start, tx->pkt_done,
3530 (int)ntohl(mgp->ss[i].fw_stats->
3531 send_done_count));
0dcffac1 3532 msleep(2000);
78ca90ea
JP
3533 netdev_info(mgp->dev, "(%d): %d %d %d %d %d %d\n",
3534 i, tx->queue_active, tx->req,
3535 tx->done, tx->pkt_start, tx->pkt_done,
3536 (int)ntohl(mgp->ss[i].fw_stats->
3537 send_done_count));
0dcffac1 3538 }
0da34b6d 3539 }
236bb5e6 3540
d0234215
BG
3541 if (!rebooted) {
3542 rtnl_lock();
3543 myri10ge_close(mgp->dev);
3544 }
0dcffac1 3545 status = myri10ge_load_firmware(mgp, 1);
0da34b6d 3546 if (status != 0)
78ca90ea 3547 netdev_err(mgp->dev, "failed to load firmware\n");
0da34b6d
BG
3548 else
3549 myri10ge_open(mgp->dev);
3550 rtnl_unlock();
3551}
3552
3553/*
3554 * We use our own timer routine rather than relying upon
3555 * netdev->tx_timeout because we have a very large hardware transmit
3556 * queue. Due to the large queue, the netdev->tx_timeout function
3557 * cannot detect a NIC with a parity error in a timely fashion if the
3558 * NIC is lightly loaded.
3559 */
3560static void myri10ge_watchdog_timer(unsigned long arg)
3561{
3562 struct myri10ge_priv *mgp;
b53bef84 3563 struct myri10ge_slice_state *ss;
d0234215 3564 int i, reset_needed, busy_slice_cnt;
626fda94 3565 u32 rx_pause_cnt;
d0234215 3566 u16 cmd;
0da34b6d
BG
3567
3568 mgp = (struct myri10ge_priv *)arg;
c7dab99b 3569
0dcffac1 3570 rx_pause_cnt = ntohl(mgp->ss[0].fw_stats->dropped_pause);
d0234215 3571 busy_slice_cnt = 0;
0dcffac1
BG
3572 for (i = 0, reset_needed = 0;
3573 i < mgp->num_slices && reset_needed == 0; ++i) {
b53bef84 3574
0dcffac1
BG
3575 ss = &mgp->ss[i];
3576 if (ss->rx_small.watchdog_needed) {
3577 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
3578 mgp->small_bytes + MXGEFW_PAD,
3579 1);
3580 if (ss->rx_small.fill_cnt - ss->rx_small.cnt >=
3581 myri10ge_fill_thresh)
3582 ss->rx_small.watchdog_needed = 0;
3583 }
3584 if (ss->rx_big.watchdog_needed) {
3585 myri10ge_alloc_rx_pages(mgp, &ss->rx_big,
3586 mgp->big_bytes, 1);
3587 if (ss->rx_big.fill_cnt - ss->rx_big.cnt >=
3588 myri10ge_fill_thresh)
3589 ss->rx_big.watchdog_needed = 0;
3590 }
3591
3592 if (ss->tx.req != ss->tx.done &&
3593 ss->tx.done == ss->watchdog_tx_done &&
3594 ss->watchdog_tx_req != ss->watchdog_tx_done) {
3595 /* nic seems like it might be stuck.. */
3596 if (rx_pause_cnt != mgp->watchdog_pause) {
3597 if (net_ratelimit())
78ca90ea
JP
3598 netdev_err(mgp->dev, "slice %d: TX paused, check link partner\n",
3599 i);
0dcffac1 3600 } else {
78ca90ea 3601 netdev_warn(mgp->dev, "slice %d stuck:", i);
0dcffac1
BG
3602 reset_needed = 1;
3603 }
626fda94 3604 }
d0234215
BG
3605 if (ss->watchdog_tx_done != ss->tx.done ||
3606 ss->watchdog_rx_done != ss->rx_done.cnt) {
3607 busy_slice_cnt++;
3608 }
0dcffac1
BG
3609 ss->watchdog_tx_done = ss->tx.done;
3610 ss->watchdog_tx_req = ss->tx.req;
d0234215
BG
3611 ss->watchdog_rx_done = ss->rx_done.cnt;
3612 }
3613 /* if we've sent or received no traffic, poll the NIC to
3614 * ensure it is still there. Otherwise, we risk not noticing
3615 * an error in a timely fashion */
3616 if (busy_slice_cnt == 0) {
3617 pci_read_config_word(mgp->pdev, PCI_COMMAND, &cmd);
3618 if ((cmd & PCI_COMMAND_MASTER) == 0) {
3619 reset_needed = 1;
3620 }
626fda94 3621 }
626fda94 3622 mgp->watchdog_pause = rx_pause_cnt;
0dcffac1
BG
3623
3624 if (reset_needed) {
3625 schedule_work(&mgp->watchdog_work);
3626 } else {
3627 /* rearm timer */
3628 mod_timer(&mgp->watchdog_timer,
3629 jiffies + myri10ge_watchdog_timeout * HZ);
3630 }
0da34b6d
BG
3631}
3632
77929732
BG
3633static void myri10ge_free_slices(struct myri10ge_priv *mgp)
3634{
3635 struct myri10ge_slice_state *ss;
3636 struct pci_dev *pdev = mgp->pdev;
3637 size_t bytes;
3638 int i;
3639
3640 if (mgp->ss == NULL)
3641 return;
3642
3643 for (i = 0; i < mgp->num_slices; i++) {
3644 ss = &mgp->ss[i];
3645 if (ss->rx_done.entry != NULL) {
3646 bytes = mgp->max_intr_slots *
3647 sizeof(*ss->rx_done.entry);
3648 dma_free_coherent(&pdev->dev, bytes,
3649 ss->rx_done.entry, ss->rx_done.bus);
3650 ss->rx_done.entry = NULL;
3651 }
3652 if (ss->fw_stats != NULL) {
3653 bytes = sizeof(*ss->fw_stats);
3654 dma_free_coherent(&pdev->dev, bytes,
3655 ss->fw_stats, ss->fw_stats_bus);
3656 ss->fw_stats = NULL;
cda6587c 3657 netif_napi_del(&ss->napi);
77929732
BG
3658 }
3659 }
3660 kfree(mgp->ss);
3661 mgp->ss = NULL;
3662}
3663
3664static int myri10ge_alloc_slices(struct myri10ge_priv *mgp)
3665{
3666 struct myri10ge_slice_state *ss;
3667 struct pci_dev *pdev = mgp->pdev;
3668 size_t bytes;
3669 int i;
3670
3671 bytes = sizeof(*mgp->ss) * mgp->num_slices;
3672 mgp->ss = kzalloc(bytes, GFP_KERNEL);
3673 if (mgp->ss == NULL) {
3674 return -ENOMEM;
3675 }
3676
3677 for (i = 0; i < mgp->num_slices; i++) {
3678 ss = &mgp->ss[i];
3679 bytes = mgp->max_intr_slots * sizeof(*ss->rx_done.entry);
3680 ss->rx_done.entry = dma_alloc_coherent(&pdev->dev, bytes,
3681 &ss->rx_done.bus,
3682 GFP_KERNEL);
3683 if (ss->rx_done.entry == NULL)
3684 goto abort;
3685 memset(ss->rx_done.entry, 0, bytes);
3686 bytes = sizeof(*ss->fw_stats);
3687 ss->fw_stats = dma_alloc_coherent(&pdev->dev, bytes,
3688 &ss->fw_stats_bus,
3689 GFP_KERNEL);
3690 if (ss->fw_stats == NULL)
3691 goto abort;
3692 ss->mgp = mgp;
3693 ss->dev = mgp->dev;
3694 netif_napi_add(ss->dev, &ss->napi, myri10ge_poll,
3695 myri10ge_napi_weight);
3696 }
3697 return 0;
3698abort:
3699 myri10ge_free_slices(mgp);
3700 return -ENOMEM;
3701}
3702
3703/*
3704 * This function determines the number of slices supported.
25985edc 3705 * The number slices is the minimum of the number of CPUS,
77929732
BG
3706 * the number of MSI-X irqs supported, the number of slices
3707 * supported by the firmware
3708 */
3709static void myri10ge_probe_slices(struct myri10ge_priv *mgp)
3710{
3711 struct myri10ge_cmd cmd;
3712 struct pci_dev *pdev = mgp->pdev;
3713 char *old_fw;
7d351035 3714 bool old_allocated;
77929732
BG
3715 int i, status, ncpus, msix_cap;
3716
3717 mgp->num_slices = 1;
3718 msix_cap = pci_find_capability(pdev, PCI_CAP_ID_MSIX);
3719 ncpus = num_online_cpus();
3720
3721 if (myri10ge_max_slices == 1 || msix_cap == 0 ||
3722 (myri10ge_max_slices == -1 && ncpus < 2))
3723 return;
3724
3725 /* try to load the slice aware rss firmware */
3726 old_fw = mgp->fw_name;
7d351035
RR
3727 old_allocated = mgp->fw_name_allocated;
3728 /* don't free old_fw if we override it. */
3729 mgp->fw_name_allocated = false;
3730
13b2738c
BG
3731 if (myri10ge_fw_name != NULL) {
3732 dev_info(&mgp->pdev->dev, "overriding rss firmware to %s\n",
3733 myri10ge_fw_name);
7d351035 3734 set_fw_name(mgp, myri10ge_fw_name, false);
13b2738c 3735 } else if (old_fw == myri10ge_fw_aligned)
7d351035 3736 set_fw_name(mgp, myri10ge_fw_rss_aligned, false);
77929732 3737 else
7d351035 3738 set_fw_name(mgp, myri10ge_fw_rss_unaligned, false);
77929732
BG
3739 status = myri10ge_load_firmware(mgp, 0);
3740 if (status != 0) {
3741 dev_info(&pdev->dev, "Rss firmware not found\n");
7d351035
RR
3742 if (old_allocated)
3743 kfree(old_fw);
77929732
BG
3744 return;
3745 }
3746
3747 /* hit the board with a reset to ensure it is alive */
3748 memset(&cmd, 0, sizeof(cmd));
3749 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_RESET, &cmd, 0);
3750 if (status != 0) {
3751 dev_err(&mgp->pdev->dev, "failed reset\n");
3752 goto abort_with_fw;
77929732
BG
3753 }
3754
3755 mgp->max_intr_slots = cmd.data0 / sizeof(struct mcp_slot);
3756
3757 /* tell it the size of the interrupt queues */
3758 cmd.data0 = mgp->max_intr_slots * sizeof(struct mcp_slot);
3759 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_SIZE, &cmd, 0);
3760 if (status != 0) {
3761 dev_err(&mgp->pdev->dev, "failed MXGEFW_CMD_SET_INTRQ_SIZE\n");
3762 goto abort_with_fw;
3763 }
3764
3765 /* ask the maximum number of slices it supports */
3766 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_RSS_QUEUES, &cmd, 0);
3767 if (status != 0)
3768 goto abort_with_fw;
3769 else
3770 mgp->num_slices = cmd.data0;
3771
3772 /* Only allow multiple slices if MSI-X is usable */
3773 if (!myri10ge_msi) {
3774 goto abort_with_fw;
3775 }
3776
3777 /* if the admin did not specify a limit to how many
3778 * slices we should use, cap it automatically to the
3779 * number of CPUs currently online */
3780 if (myri10ge_max_slices == -1)
3781 myri10ge_max_slices = ncpus;
3782
3783 if (mgp->num_slices > myri10ge_max_slices)
3784 mgp->num_slices = myri10ge_max_slices;
3785
3786 /* Now try to allocate as many MSI-X vectors as we have
3787 * slices. We give up on MSI-X if we can only get a single
3788 * vector. */
3789
baeb2ffa
JP
3790 mgp->msix_vectors = kcalloc(mgp->num_slices, sizeof(*mgp->msix_vectors),
3791 GFP_KERNEL);
77929732
BG
3792 if (mgp->msix_vectors == NULL)
3793 goto disable_msix;
3794 for (i = 0; i < mgp->num_slices; i++) {
3795 mgp->msix_vectors[i].entry = i;
3796 }
3797
3798 while (mgp->num_slices > 1) {
3799 /* make sure it is a power of two */
3800 while (!is_power_of_2(mgp->num_slices))
3801 mgp->num_slices--;
3802 if (mgp->num_slices == 1)
3803 goto disable_msix;
3804 status = pci_enable_msix(pdev, mgp->msix_vectors,
3805 mgp->num_slices);
3806 if (status == 0) {
3807 pci_disable_msix(pdev);
7d351035
RR
3808 if (old_allocated)
3809 kfree(old_fw);
77929732
BG
3810 return;
3811 }
3812 if (status > 0)
3813 mgp->num_slices = status;
3814 else
3815 goto disable_msix;
3816 }
3817
3818disable_msix:
3819 if (mgp->msix_vectors != NULL) {
3820 kfree(mgp->msix_vectors);
3821 mgp->msix_vectors = NULL;
3822 }
3823
3824abort_with_fw:
3825 mgp->num_slices = 1;
7d351035 3826 set_fw_name(mgp, old_fw, old_allocated);
77929732
BG
3827 myri10ge_load_firmware(mgp, 0);
3828}
77929732 3829
8126089f
SH
3830static const struct net_device_ops myri10ge_netdev_ops = {
3831 .ndo_open = myri10ge_open,
3832 .ndo_stop = myri10ge_close,
3833 .ndo_start_xmit = myri10ge_xmit,
3834 .ndo_get_stats = myri10ge_get_stats,
3835 .ndo_validate_addr = eth_validate_addr,
3836 .ndo_change_mtu = myri10ge_change_mtu,
3837 .ndo_set_multicast_list = myri10ge_set_multicast_list,
3838 .ndo_set_mac_address = myri10ge_set_mac_address,
3839};
3840
0da34b6d
BG
3841static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
3842{
3843 struct net_device *netdev;
3844 struct myri10ge_priv *mgp;
3845 struct device *dev = &pdev->dev;
0da34b6d
BG
3846 int i;
3847 int status = -ENXIO;
0da34b6d 3848 int dac_enabled;
00b5e505 3849 unsigned hdr_offset, ss_offset;
2d90b0aa 3850 static int board_number;
0da34b6d 3851
236bb5e6 3852 netdev = alloc_etherdev_mq(sizeof(*mgp), MYRI10GE_MAX_SLICES);
0da34b6d
BG
3853 if (netdev == NULL) {
3854 dev_err(dev, "Could not allocate ethernet device\n");
3855 return -ENOMEM;
3856 }
3857
b245fb67
MH
3858 SET_NETDEV_DEV(netdev, &pdev->dev);
3859
0da34b6d 3860 mgp = netdev_priv(netdev);
0da34b6d
BG
3861 mgp->dev = netdev;
3862 mgp->pdev = pdev;
3863 mgp->csum_flag = MXGEFW_FLAGS_CKSUM;
3864 mgp->pause = myri10ge_flow_control;
3865 mgp->intr_coal_delay = myri10ge_intr_coal_delay;
c58ac5ca 3866 mgp->msg_enable = netif_msg_init(myri10ge_debug, MYRI10GE_MSG_DEFAULT);
2d90b0aa 3867 mgp->board_number = board_number;
0da34b6d
BG
3868 init_waitqueue_head(&mgp->down_wq);
3869
3870 if (pci_enable_device(pdev)) {
3871 dev_err(&pdev->dev, "pci_enable_device call failed\n");
3872 status = -ENODEV;
3873 goto abort_with_netdev;
3874 }
0da34b6d
BG
3875
3876 /* Find the vendor-specific cap so we can check
3877 * the reboot register later on */
3878 mgp->vendor_specific_offset
3879 = pci_find_capability(pdev, PCI_CAP_ID_VNDR);
3880
3881 /* Set our max read request to 4KB */
302d242c 3882 status = pcie_set_readrq(pdev, 4096);
0da34b6d
BG
3883 if (status != 0) {
3884 dev_err(&pdev->dev, "Error %d writing PCI_EXP_DEVCTL\n",
3885 status);
e3fd5534 3886 goto abort_with_enabled;
0da34b6d
BG
3887 }
3888
3889 pci_set_master(pdev);
3890 dac_enabled = 1;
6a35528a 3891 status = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
0da34b6d
BG
3892 if (status != 0) {
3893 dac_enabled = 0;
3894 dev_err(&pdev->dev,
898eb71c
JP
3895 "64-bit pci address mask was refused, "
3896 "trying 32-bit\n");
284901a9 3897 status = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
0da34b6d
BG
3898 }
3899 if (status != 0) {
3900 dev_err(&pdev->dev, "Error %d setting DMA mask\n", status);
e3fd5534 3901 goto abort_with_enabled;
0da34b6d 3902 }
6a35528a 3903 (void)pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
b10c0668
BG
3904 mgp->cmd = dma_alloc_coherent(&pdev->dev, sizeof(*mgp->cmd),
3905 &mgp->cmd_bus, GFP_KERNEL);
0da34b6d 3906 if (mgp->cmd == NULL)
e3fd5534 3907 goto abort_with_enabled;
0da34b6d 3908
0da34b6d
BG
3909 mgp->board_span = pci_resource_len(pdev, 0);
3910 mgp->iomem_base = pci_resource_start(pdev, 0);
3911 mgp->mtrr = -1;
276e26c3 3912 mgp->wc_enabled = 0;
0da34b6d
BG
3913#ifdef CONFIG_MTRR
3914 mgp->mtrr = mtrr_add(mgp->iomem_base, mgp->board_span,
3915 MTRR_TYPE_WRCOMB, 1);
276e26c3
BG
3916 if (mgp->mtrr >= 0)
3917 mgp->wc_enabled = 1;
0da34b6d 3918#endif
c7f80993 3919 mgp->sram = ioremap_wc(mgp->iomem_base, mgp->board_span);
0da34b6d
BG
3920 if (mgp->sram == NULL) {
3921 dev_err(&pdev->dev, "ioremap failed for %ld bytes at 0x%lx\n",
3922 mgp->board_span, mgp->iomem_base);
3923 status = -ENXIO;
c7f80993 3924 goto abort_with_mtrr;
0da34b6d 3925 }
00b5e505
BG
3926 hdr_offset =
3927 ntohl(__raw_readl(mgp->sram + MCP_HEADER_PTR_OFFSET)) & 0xffffc;
3928 ss_offset = hdr_offset + offsetof(struct mcp_gen_header, string_specs);
3929 mgp->sram_size = ntohl(__raw_readl(mgp->sram + ss_offset));
3930 if (mgp->sram_size > mgp->board_span ||
3931 mgp->sram_size <= MYRI10GE_FW_OFFSET) {
3932 dev_err(&pdev->dev,
3933 "invalid sram_size %dB or board span %ldB\n",
3934 mgp->sram_size, mgp->board_span);
3935 goto abort_with_ioremap;
3936 }
0da34b6d 3937 memcpy_fromio(mgp->eeprom_strings,
00b5e505 3938 mgp->sram + mgp->sram_size, MYRI10GE_EEPROM_STRINGS_SIZE);
0da34b6d
BG
3939 memset(mgp->eeprom_strings + MYRI10GE_EEPROM_STRINGS_SIZE - 2, 0, 2);
3940 status = myri10ge_read_mac_addr(mgp);
3941 if (status)
3942 goto abort_with_ioremap;
3943
3944 for (i = 0; i < ETH_ALEN; i++)
3945 netdev->dev_addr[i] = mgp->mac_addr[i];
3946
5443e9ea
BG
3947 myri10ge_select_firmware(mgp);
3948
0dcffac1 3949 status = myri10ge_load_firmware(mgp, 1);
0da34b6d
BG
3950 if (status != 0) {
3951 dev_err(&pdev->dev, "failed to load firmware\n");
0dcffac1
BG
3952 goto abort_with_ioremap;
3953 }
3954 myri10ge_probe_slices(mgp);
3955 status = myri10ge_alloc_slices(mgp);
3956 if (status != 0) {
3957 dev_err(&pdev->dev, "failed to alloc slice state\n");
3958 goto abort_with_firmware;
0da34b6d 3959 }
c9920268
BH
3960 netif_set_real_num_tx_queues(netdev, mgp->num_slices);
3961 netif_set_real_num_rx_queues(netdev, mgp->num_slices);
0da34b6d
BG
3962 status = myri10ge_reset(mgp);
3963 if (status != 0) {
3964 dev_err(&pdev->dev, "failed reset\n");
0dcffac1 3965 goto abort_with_slices;
0da34b6d 3966 }
5dd2d332 3967#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
3968 myri10ge_setup_dca(mgp);
3969#endif
0da34b6d
BG
3970 pci_set_drvdata(pdev, mgp);
3971 if ((myri10ge_initial_mtu + ETH_HLEN) > MYRI10GE_MAX_ETHER_MTU)
3972 myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
3973 if ((myri10ge_initial_mtu + ETH_HLEN) < 68)
3974 myri10ge_initial_mtu = 68;
8126089f
SH
3975
3976 netdev->netdev_ops = &myri10ge_netdev_ops;
0da34b6d 3977 netdev->mtu = myri10ge_initial_mtu;
0da34b6d 3978 netdev->base_addr = mgp->iomem_base;
4f93fde0 3979 netdev->features = mgp->features;
236bb5e6 3980
0da34b6d
BG
3981 if (dac_enabled)
3982 netdev->features |= NETIF_F_HIGHDMA;
2552c31b 3983 netdev->features |= NETIF_F_LRO;
0da34b6d 3984
dddc045e
BG
3985 netdev->vlan_features |= mgp->features;
3986 if (mgp->fw_ver_tiny < 37)
3987 netdev->vlan_features &= ~NETIF_F_TSO6;
3988 if (mgp->fw_ver_tiny < 32)
3989 netdev->vlan_features &= ~NETIF_F_TSO;
3990
21d05db1
BG
3991 /* make sure we can get an irq, and that MSI can be
3992 * setup (if available). Also ensure netdev->irq
3993 * is set to correct value if MSI is enabled */
3994 status = myri10ge_request_irq(mgp);
3995 if (status != 0)
3996 goto abort_with_firmware;
3997 netdev->irq = pdev->irq;
3998 myri10ge_free_irq(mgp);
3999
0da34b6d
BG
4000 /* Save configuration space to be restored if the
4001 * nic resets due to a parity error */
83f6e152 4002 pci_save_state(pdev);
0da34b6d
BG
4003
4004 /* Setup the watchdog timer */
4005 setup_timer(&mgp->watchdog_timer, myri10ge_watchdog_timer,
4006 (unsigned long)mgp);
4007
59081825 4008 spin_lock_init(&mgp->stats_lock);
0da34b6d 4009 SET_ETHTOOL_OPS(netdev, &myri10ge_ethtool_ops);
c4028958 4010 INIT_WORK(&mgp->watchdog_work, myri10ge_watchdog);
0da34b6d
BG
4011 status = register_netdev(netdev);
4012 if (status != 0) {
4013 dev_err(&pdev->dev, "register_netdev failed: %d\n", status);
7adda30c 4014 goto abort_with_state;
0da34b6d 4015 }
0dcffac1
BG
4016 if (mgp->msix_enabled)
4017 dev_info(dev, "%d MSI-X IRQs, tx bndry %d, fw %s, WC %s\n",
4018 mgp->num_slices, mgp->tx_boundary, mgp->fw_name,
4019 (mgp->wc_enabled ? "Enabled" : "Disabled"));
4020 else
4021 dev_info(dev, "%s IRQ %d, tx bndry %d, fw %s, WC %s\n",
4022 mgp->msi_enabled ? "MSI" : "xPIC",
4023 netdev->irq, mgp->tx_boundary, mgp->fw_name,
4024 (mgp->wc_enabled ? "Enabled" : "Disabled"));
0da34b6d 4025
2d90b0aa 4026 board_number++;
0da34b6d
BG
4027 return 0;
4028
7adda30c 4029abort_with_state:
83f6e152 4030 pci_restore_state(pdev);
0da34b6d 4031
0dcffac1
BG
4032abort_with_slices:
4033 myri10ge_free_slices(mgp);
4034
0da34b6d
BG
4035abort_with_firmware:
4036 myri10ge_dummy_rdma(mgp, 0);
4037
0da34b6d 4038abort_with_ioremap:
0f840011
BG
4039 if (mgp->mac_addr_string != NULL)
4040 dev_err(&pdev->dev,
4041 "myri10ge_probe() failed: MAC=%s, SN=%ld\n",
4042 mgp->mac_addr_string, mgp->serial_number);
0da34b6d
BG
4043 iounmap(mgp->sram);
4044
c7f80993 4045abort_with_mtrr:
0da34b6d
BG
4046#ifdef CONFIG_MTRR
4047 if (mgp->mtrr >= 0)
4048 mtrr_del(mgp->mtrr, mgp->iomem_base, mgp->board_span);
4049#endif
b10c0668
BG
4050 dma_free_coherent(&pdev->dev, sizeof(*mgp->cmd),
4051 mgp->cmd, mgp->cmd_bus);
0da34b6d 4052
e3fd5534
BG
4053abort_with_enabled:
4054 pci_disable_device(pdev);
0da34b6d 4055
e3fd5534 4056abort_with_netdev:
7d351035 4057 set_fw_name(mgp, NULL, false);
0da34b6d
BG
4058 free_netdev(netdev);
4059 return status;
4060}
4061
4062/*
4063 * myri10ge_remove
4064 *
4065 * Does what is necessary to shutdown one Myrinet device. Called
4066 * once for each Myrinet card by the kernel when a module is
4067 * unloaded.
4068 */
4069static void myri10ge_remove(struct pci_dev *pdev)
4070{
4071 struct myri10ge_priv *mgp;
4072 struct net_device *netdev;
0da34b6d
BG
4073
4074 mgp = pci_get_drvdata(pdev);
4075 if (mgp == NULL)
4076 return;
4077
23f333a2 4078 cancel_work_sync(&mgp->watchdog_work);
0da34b6d
BG
4079 netdev = mgp->dev;
4080 unregister_netdev(netdev);
0da34b6d 4081
5dd2d332 4082#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
4083 myri10ge_teardown_dca(mgp);
4084#endif
0da34b6d
BG
4085 myri10ge_dummy_rdma(mgp, 0);
4086
7adda30c 4087 /* avoid a memory leak */
83f6e152 4088 pci_restore_state(pdev);
7adda30c 4089
0da34b6d
BG
4090 iounmap(mgp->sram);
4091
4092#ifdef CONFIG_MTRR
4093 if (mgp->mtrr >= 0)
4094 mtrr_del(mgp->mtrr, mgp->iomem_base, mgp->board_span);
4095#endif
0dcffac1
BG
4096 myri10ge_free_slices(mgp);
4097 if (mgp->msix_vectors != NULL)
4098 kfree(mgp->msix_vectors);
b10c0668
BG
4099 dma_free_coherent(&pdev->dev, sizeof(*mgp->cmd),
4100 mgp->cmd, mgp->cmd_bus);
0da34b6d 4101
7d351035 4102 set_fw_name(mgp, NULL, false);
0da34b6d 4103 free_netdev(netdev);
e3fd5534 4104 pci_disable_device(pdev);
0da34b6d
BG
4105 pci_set_drvdata(pdev, NULL);
4106}
4107
b10c0668 4108#define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E 0x0008
a07bc1ff 4109#define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9 0x0009
0da34b6d 4110
a3aa1884 4111static DEFINE_PCI_DEVICE_TABLE(myri10ge_pci_tbl) = {
b10c0668 4112 {PCI_DEVICE(PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E)},
a07bc1ff
BG
4113 {PCI_DEVICE
4114 (PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9)},
0da34b6d
BG
4115 {0},
4116};
4117
97131079
BG
4118MODULE_DEVICE_TABLE(pci, myri10ge_pci_tbl);
4119
0da34b6d
BG
4120static struct pci_driver myri10ge_driver = {
4121 .name = "myri10ge",
4122 .probe = myri10ge_probe,
4123 .remove = myri10ge_remove,
4124 .id_table = myri10ge_pci_tbl,
4125#ifdef CONFIG_PM
4126 .suspend = myri10ge_suspend,
4127 .resume = myri10ge_resume,
4128#endif
4129};
4130
5dd2d332 4131#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
4132static int
4133myri10ge_notify_dca(struct notifier_block *nb, unsigned long event, void *p)
4134{
4135 int err = driver_for_each_device(&myri10ge_driver.driver,
4136 NULL, &event,
4137 myri10ge_notify_dca_device);
4138
4139 if (err)
4140 return NOTIFY_BAD;
4141 return NOTIFY_DONE;
4142}
4143
4144static struct notifier_block myri10ge_dca_notifier = {
4145 .notifier_call = myri10ge_notify_dca,
4146 .next = NULL,
4147 .priority = 0,
4148};
4ee2ac51 4149#endif /* CONFIG_MYRI10GE_DCA */
981813d8 4150
0da34b6d
BG
4151static __init int myri10ge_init_module(void)
4152{
78ca90ea 4153 pr_info("Version %s\n", MYRI10GE_VERSION_STR);
0dcffac1 4154
236bb5e6 4155 if (myri10ge_rss_hash > MXGEFW_RSS_HASH_TYPE_MAX) {
78ca90ea
JP
4156 pr_err("Illegal rssh hash type %d, defaulting to source port\n",
4157 myri10ge_rss_hash);
0dcffac1
BG
4158 myri10ge_rss_hash = MXGEFW_RSS_HASH_TYPE_SRC_PORT;
4159 }
5dd2d332 4160#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
4161 dca_register_notify(&myri10ge_dca_notifier);
4162#endif
236bb5e6
BG
4163 if (myri10ge_max_slices > MYRI10GE_MAX_SLICES)
4164 myri10ge_max_slices = MYRI10GE_MAX_SLICES;
0dcffac1 4165
0da34b6d
BG
4166 return pci_register_driver(&myri10ge_driver);
4167}
4168
4169module_init(myri10ge_init_module);
4170
4171static __exit void myri10ge_cleanup_module(void)
4172{
5dd2d332 4173#ifdef CONFIG_MYRI10GE_DCA
981813d8
BG
4174 dca_unregister_notify(&myri10ge_dca_notifier);
4175#endif
0da34b6d
BG
4176 pci_unregister_driver(&myri10ge_driver);
4177}
4178
4179module_exit(myri10ge_cleanup_module);